<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Leena Haque on Medium]]></title>
        <description><![CDATA[Stories by Leena Haque on Medium]]></description>
        <link>https://medium.com/@leenahaque1?source=rss-4fa29c67eccc------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*wAkQpyQVq-a4_Nn9BubONA@2x.jpeg</url>
            <title>Stories by Leena Haque on Medium</title>
            <link>https://medium.com/@leenahaque1?source=rss-4fa29c67eccc------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 06:24:13 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@leenahaque1/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Accessibility vs Inclusion: Are We Asking the Right Questions About Autism Support?]]></title>
            <link>https://medium.com/@leenahaque1/accessibility-vs-inclusion-are-we-asking-the-right-questions-about-autism-support-9a4cacb4a35d?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/9a4cacb4a35d</guid>
            <category><![CDATA[autism]]></category>
            <category><![CDATA[neurodiversity]]></category>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[disability]]></category>
            <category><![CDATA[assistive-technology]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Fri, 23 Oct 2026 11:12:45 GMT</pubDate>
            <atom:updated>2026-10-23T11:12:45.172Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Image titled “Accessibility vs Inclusion: Why Different Autism Support Needs Matter.” A South Asian person with a backpack faces colourful pathways and pastel signposts reading Access, Support, Inclusion and Opportunity, surrounded by accessibility symbols in muted fluorescent colours." src="https://cdn-images-1.medium.com/max/856/1*_m-3ECOQ4NSqeRTDBDdz6A.png" /></figure><p>Imagine an organisation introduces an online appointment system designed to make its services more accessible.</p><p>For an autistic person who finds telephone conversations difficult, this could be a welcome improvement. They can book an appointment without the anxiety or communication difficulties that a phone call might involve.</p><p>But someone else may struggle with online forms, processing written information or navigating an unfamiliar website.</p><p>The same improvement could make accessing the service easier for one person while creating difficulties for another, particularly if alternative booking methods are removed.</p><p>Neither person’s needs are less important.</p><p>So when we describe something as accessible, what are we actually measuring?</p><p>Are we looking at the barriers we’ve removed, or are we also considering those that remain?</p><p>I think this distinction deserves more attention in conversations about autism, neurodiversity and disability inclusion.</p><p>This isn’t about questioning anyone’s lived experience or suggesting that existing accessibility initiatives have little value. Improvements that make people’s lives easier should be recognised.</p><p>But perhaps we also need to become more curious about who benefits, who might still face difficulties and what those differences can teach us.</p><p><strong>Accessibility and inclusion: understanding the difference</strong></p><p>Accessibility and inclusion are closely connected, but they’re not interchangeable.</p><p>Accessibility generally focuses on identifying and reducing barriers that prevent people from using services, environments or information.</p><p>Inclusion considers whether people with different needs can participate meaningfully and have their preferences respected.</p><p>An accessibility improvement can therefore be successful without meeting everyone’s needs.</p><p>Consider a workplace that introduces a quiet room for employees who experience sensory overload.</p><p>For some autistic employees, this could provide valuable relief from a busy working environment.</p><p>Others might benefit more from changes to lighting, clearer communication, predictable routines or flexible working arrangements.</p><p>The quiet room is still a positive development. Its value doesn’t disappear simply because it isn’t suitable for everyone.</p><p>But recognising what it doesn’t address can help an organisation understand where further adjustments might be needed.</p><p>Perhaps one of the challenges with accessibility is that we often recognise progress through the barriers we’ve successfully removed rather than the barriers that still exist.</p><p>An improvement can be genuinely valuable without being universally effective.</p><p>Recognising its limitations doesn’t diminish its success. It helps us understand what needs to happen next.</p><p><strong>What can lived experience tell us about autism support?</strong></p><p>Lived experience is essential to understanding autism.</p><p>Autistic people can identify difficulties, preferences and everyday barriers that may not be obvious to people who haven’t encountered them.</p><p>Their contributions have helped challenge misconceptions and improve conversations about disability.</p><p>However, personal experiences also remind us how varied autism can be.</p><p>Autistic people differ in their communication styles, sensory experiences, abilities, preferences and support needs.</p><p>Some live independently, while others require substantial assistance with everyday activities. Some communicate through speech, while others use augmentative and alternative communication (AAC) or other methods.</p><p>Two people may encounter the same environment but experience very different challenges.</p><p>This raises an interesting question.</p><p><strong>When we learn that an approach has worked well for someone, how do we decide whether it might also work for others?</strong></p><p>A personal success story can show us what is possible. It may inspire an organisation to introduce an adjustment that benefits many people.</p><p>But it can’t tell us everything about how that adjustment will work across a diverse population.</p><p>That is where wider consultation, research and feedback become useful.</p><p>Rather than seeing lived experience and evidence as competing forms of knowledge, we can recognise how they complement one another.</p><p>Individual experiences can help identify possibilities. Broader evaluation can help us understand their limitations.</p><p>Both contribute to better autism support.</p><p><strong>When does accessible technology become another barrier?</strong></p><p>Technology has expanded the options available to many disabled people.</p><p>Artificial intelligence, communication aids, speech-to-text tools and accessible digital services can support communication, organisation and participation in everyday activities.</p><p>For some autistic people, these developments can make a considerable difference.</p><p>But I think there’s a question worth asking whenever technology is presented as an accessibility solution.</p><p><strong>Does the technology remove a barrier, or does it change what someone needs in order to overcome that barrier?</strong></p><p>Imagine a service replaces printed information with a mobile app.</p><p>For someone who prefers digital information, this may be convenient and accessible.</p><p>Another person might find the app difficult to navigate, lack a suitable device or need assistance understanding its content.</p><p>The service has introduced a new way of accessing information, but whether it has improved inclusion depends partly on what other options remain.</p><p>This doesn’t mean technological innovation should be discouraged.</p><p>It means we should think carefully about how technology fits into the wider support available.</p><p>An accessible app may be extremely helpful. So might a telephone conversation, printed information or assistance from a trained member of staff.</p><p>Different approaches can serve different needs.</p><p>Perhaps the most useful question isn’t whether technology can replace existing methods, but whether it can expand the choices people have.</p><p><strong>Are we measuring independence or quality of life?</strong></p><p>Independence is frequently discussed as a goal of disability accessibility.</p><p>For many autistic people, being able to manage everyday activities without assistance is empowering and important.</p><p>But independence isn’t necessarily the same as autonomy, and neither should be confused with quality of life.</p><p>Someone may require ongoing support while still having considerable control over decisions affecting them.</p><p>For example, a person might need help understanding information, communicating preferences or completing practical tasks.</p><p>With appropriate assistance, they may be able to participate in decisions and activities that would otherwise be difficult to access.</p><p>Their reliance on support doesn’t make their choices less meaningful.</p><p>This makes me wonder whether we sometimes focus too heavily on how much assistance someone needs, rather than whether that assistance allows them to live in a way that suits them.</p><p>For one person, progress might mean completing a task independently.</p><p>For another, it might mean receiving consistent, reliable support that makes the task less stressful.</p><p>Both can represent improvements.</p><p><strong>Shouldn’t the measure of successful autism support be how well it responds to the person’s needs, rather than how little support they require?</strong></p><p><strong>How do we know an accessibility solution is working?</strong></p><p>When an accessibility initiative receives positive feedback, it’s natural to see that as encouraging.</p><p>But understanding its effectiveness involves more than knowing that some people have benefited.</p><p>We also need to consider the experiences of people who found it difficult to use or couldn’t use it at all.</p><p>Imagine a new communication tool is introduced and receives excellent feedback from autistic adults who are comfortable with digital technology.</p><p>That tells us something useful about its potential.</p><p>It would also be valuable to understand how the tool works for people who require substantial communication support or assistance with technology.</p><p>Were they included in testing? What difficulties did they encounter? Were alternative methods available?</p><p>These questions aren’t intended to undermine positive experiences.</p><p>They can help identify opportunities to improve the tool or provide additional options.</p><p>There is another issue worth considering.</p><p>People who find a service inaccessible may stop using it, meaning their experiences are not always captured through ordinary feedback.</p><p>If we only hear from those who successfully use an accessibility solution, we may miss important information about those who couldn’t.</p><p>This is why ongoing evaluation matters.</p><p>Accessibility isn’t necessarily something an organisation achieves once and then considers complete.</p><p>It may need to change as people’s needs, technology and services develop.</p><p><strong>Whose experiences are missing from conversations about inclusion?</strong></p><p>There are more opportunities than ever for autistic people to share their experiences publicly.</p><p>Blogs, social media, podcasts and community discussions have helped increase understanding and bring attention to issues that deserve recognition.</p><p>But participation in these conversations isn’t equally accessible to everyone.</p><p>Some people communicate comfortably through writing or public speaking. Others may need additional communication support, more time or different ways of expressing their views.</p><p>Some may choose not to share their experiences publicly.</p><p>This creates an important consideration for organisations seeking to understand autism support needs.</p><p><strong>How do we make sure consultation reaches people who may find conventional methods of participation difficult?</strong></p><p>An online survey might work well for some people but not others.</p><p>A public meeting may provide valuable discussion while remaining inaccessible to someone who experiences communication or sensory difficulties.</p><p>Individual conversations, accessible information and supported participation may help broaden the range of perspectives included.</p><p>There is also value in recognising that autistic people will not always agree with one another.</p><p>Different experiences can lead to different views about independence, technology, communication and the support people find useful.</p><p>Those differences don’t need to be resolved into one shared opinion.</p><p>They can help us understand why flexible approaches are often necessary.</p><p><strong>Why respectful disagreement can improve accessibility</strong></p><p>Conversations about disability and autism can involve deeply personal experiences.</p><p>When something has made a meaningful difference to someone’s life, it’s understandable that they may feel strongly about its value.</p><p>That experience deserves respect.</p><p>At the same time, we should be able to explore whether an approach might work differently for other people.</p><p>Someone can benefit greatly from a particular technology while another person needs a different form of support.</p><p>An organisation can introduce a valuable accessibility improvement while still having areas that need attention.</p><p>Both things can be true.</p><p>I think constructive discussion becomes easier when we recognise that questioning the wider application of a solution isn’t the same as questioning the experience of someone who benefits from it.</p><p>No individual, organisation or professional discipline can capture every aspect of autism.</p><p>Listening to different perspectives, including those that challenge our assumptions, can help us develop a more complete understanding.</p><p><strong>What would more inclusive autism support look like?</strong></p><p>I don’t believe inclusion requires finding one accessibility solution that works perfectly for everyone.</p><p>People’s needs are too varied for that to be realistic.</p><p>Instead, inclusion may be better understood as an ongoing willingness to recognise differences, provide appropriate options and respond when barriers become apparent.</p><p>For an organisation, this might mean offering both digital and non-digital ways to access services.</p><p>For an employer, it could involve providing different workplace adjustments rather than assuming one change will meet everyone’s sensory or communication needs.</p><p>For someone who requires substantial daily assistance, it may mean having reliable support that respects their preferences and helps them participate in decisions.</p><p>These approaches may look different, but they share an important principle.</p><p>They start by considering what the person needs rather than expecting everyone to use the same solution.</p><p>Of course, organisations face practical constraints, and not every adjustment will be possible in every situation.</p><p>But there is still value in asking whether alternatives have been considered and whether the people affected have had meaningful opportunities to contribute.</p><p>The goal isn’t perfection.</p><p>It’s a willingness to keep learning and improving.</p><p><strong>Accessibility vs inclusion: perhaps we need to ask different questions</strong></p><p>We’ve made encouraging progress in autism awareness, accessible technology and disability inclusion.</p><p>People are sharing experiences, organisations are introducing adjustments, and barriers that once received little attention are being recognised.</p><p>That progress matters.</p><p>But perhaps we need to think more carefully about how we evaluate it.</p><p>Instead of asking only whether an accessibility solution works, we could also ask who it works for.</p><p>Instead of measuring success solely through independence, we could consider choice, dignity and quality of life.</p><p>And rather than assuming that positive feedback tells us everything we need to know, we could pay attention to the people who may still be encountering difficulties.</p><p>None of this requires us to dismiss the progress already being made.</p><p>It simply encourages us to build on it.</p><p>For me, accessibility and inclusion should work alongside one another.</p><p>Accessibility helps us identify and reduce barriers. Inclusion encourages us to consider the people who may still be facing them.</p><p><strong>Perhaps genuine inclusion isn’t about finding a perfect solution. It’s about remaining willing to ask whose needs we haven’t yet fully understood.</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9a4cacb4a35d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Should We Evaluate New Ideas in Cognitive Accessibility?]]></title>
            <link>https://medium.com/@leenahaque1/how-should-we-evaluate-new-ideas-in-cognitive-accessibility-17b20c8eaec8?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/17b20c8eaec8</guid>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ux-research]]></category>
            <category><![CDATA[inclusive-design]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Tue, 20 Oct 2026 01:26:38 GMT</pubDate>
            <atom:updated>2026-10-20T01:26:38.762Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Editorial collage featuring a thoughtful woman of colour, a wheelchair user on an accessible pathway, and layered architectural and abstract shapes representing inclusive design and cognitive accessibility. The headline reads: “How Should We Evaluate New Ideas in Cognitive Accessibility?”" src="https://cdn-images-1.medium.com/max/856/1*dO2iLh5fFFEwSeZTyB51UA.png" /></figure><p><em>On research originality, professional expertise and the evidence needed to distinguish meaningful innovation from familiar concepts.</em></p><p>Cognitive accessibility presents an unusual challenge for researchers and practitioners. The problems it addresses are well established, yet the ways we understand and respond to them continue to develop.</p><p>We know that attention, memory, language, comprehension and executive functioning can shape how people interact with digital systems. We also know that these experiences vary considerably between individuals and contexts.</p><p>The World Wide Web Consortium’s work on cognitive and learning disabilities has made an important contribution to understanding these barriers. Its guidance provides practical recommendations for making digital content and interfaces easier to understand and use.</p><p>Yet established guidance doesn’t mean every question has been answered.</p><p>New technologies create new interactions. Research disciplines develop different ways of describing familiar problems. Ideas that originated in one field may prove useful when examined alongside those from another.</p><p>This raises a question that deserves more attention in accessibility research: <strong>how should we decide whether an emerging framework makes a worthwhile contribution?</strong></p><p>It’s tempting to answer through familiarity. If an idea resembles existing research, it may be dismissed as unnecessary. If it introduces unfamiliar language, it may be regarded with suspicion. If it comes from an unexpected institution or discipline, its legitimacy may be questioned before its claims have been examined.</p><p>None of these reactions tells us very much about the quality of the work.</p><p>A more useful approach is to evaluate emerging frameworks against clear intellectual and practical standards.</p><p><strong>1. Conceptual clarity: What is the framework actually proposing?</strong></p><p>Before assessing originality, we need to understand what a framework claims to explain.</p><p>Some frameworks introduce genuinely new concepts. Others reorganise existing knowledge, identify relationships between established ideas or provide a different way of communicating complex subjects.</p><p>These are different kinds of contribution.</p><p>A useful evaluation begins with the definitions.</p><p>Are the key concepts explained clearly? Are the boundaries of the framework understandable? Does it distinguish between observations, hypotheses and recommendations?</p><p>Terminology matters here, but unfamiliar language isn’t necessarily a weakness.</p><p>A new term might help describe a relationship that existing vocabulary doesn’t capture conveniently. It might also obscure a familiar idea behind an unnecessary label.</p><p>The distinction is established by examining the term’s meaning and usefulness.</p><p>A framework shouldn’t receive credit merely for naming something. Equally, introducing vocabulary isn’t evidence of an attempt to claim intellectual ownership over the underlying subject.</p><p>The appropriate question is whether the language improves understanding.</p><p><strong>2. Relationship to prior research: What does it add?</strong></p><p>Originality is often discussed as though it means discovering something nobody has previously considered.</p><p>In practice, research contributions take several forms.</p><p>A study may produce new empirical findings. A theoretical framework may bring previously separate concepts into a coherent relationship. An applied model may help practitioners use established knowledge in a different setting.</p><p>The National Academies’ <em>Facilitating Interdisciplinary Research</em> examines the importance of integrating knowledge and methods across disciplinary boundaries.</p><p>This is particularly relevant to cognitive accessibility, which draws upon areas including psychology, disability studies, human-computer interaction, education and inclusive design.</p><p>Consider the relationship between cognitive load and digital interaction.</p><p>Cognitive load has an established research history, as do many principles used to make interfaces easier to understand. But applying those ideas to emerging technologies, different interaction patterns or organisational decision-making may raise questions that existing guidance hasn’t fully addressed.</p><p>That doesn’t mean every proposed synthesis is valuable.</p><p>It means its value needs to be examined rather than inferred from the age of its component ideas.</p><p>A rigorous review should identify relevant predecessors, explain the similarities and establish whether the proposed work contributes a useful distinction, application or synthesis.</p><p>Finding earlier scholarship is the beginning of that assessment.</p><p>It isn’t necessarily the conclusion.</p><p><strong>3. Evidential support: What can we reasonably conclude?</strong></p><p>Conceptual frameworks and empirical findings shouldn’t be evaluated as though they make identical claims.</p><p>An exploratory framework may organise questions for future investigation. An empirical study may test a defined hypothesis. A practical intervention may claim to improve measurable outcomes.</p><p>Each requires evidence appropriate to its purpose.</p><p>For an exploratory proposal, we might ask whether its assumptions are plausible, whether its concepts are coherent and whether it identifies questions worth investigating.</p><p>For an intervention, we’d expect stronger evidence about effectiveness, limitations and potential unintended consequences.</p><p>Confusing these stages can distort professional debate.</p><p>An early-stage framework may not yet demonstrate practical effectiveness. That’s a limitation worth acknowledging. It doesn’t automatically establish that the framework is intellectually worthless.</p><p>Similarly, identifying a possible risk isn’t the same as demonstrating actual harm.</p><p>Accessibility has good reasons to take harm seriously. Ineffective products and poorly designed interventions can create barriers, waste resources or give organisations false confidence that accessibility problems have been resolved.</p><p>But claims about harm need precision.</p><p>A useful assessment identifies the people potentially affected, the mechanism through which harm might arise and the evidence supporting the concern.</p><p>Sometimes the evidence will justify a strong warning. Sometimes it will support a more limited conclusion or a recommendation for further investigation.</p><p>The strength of the language should follow the strength of the evidence.</p><p><strong>4. Practical usefulness: Does the work help anyone make better decisions?</strong></p><p>Not every worthwhile research contribution produces an immediately deployable tool.</p><p>Some work is valuable because it changes how a problem is understood. Other work provides a method for testing ideas or helps practitioners identify questions they’ve previously overlooked.</p><p>Nevertheless, practical relevance matters in accessibility.</p><p>A framework should eventually be able to explain what it helps people do.</p><p>Does it support more informed design decisions? Does it help researchers identify gaps? Can it improve how organisations understand cognitive barriers? Does it make relevant knowledge easier to communicate across disciplines?</p><p>These questions are particularly important when new terminology or conceptual models are introduced.</p><p>If a framework merely restates familiar principles without adding clarity or utility, that weakness should be recognised.</p><p>But usefulness shouldn’t be defined so narrowly that only established methods qualify.</p><p>Emerging technologies frequently create situations that existing guidance doesn’t address in sufficient detail. Researchers may need to develop provisional models before they can test their practical value.</p><p>The important distinction is between proposing a useful line of inquiry and claiming to have already demonstrated a solution.</p><p>Clear communication about that distinction allows readers to judge the work on its actual merits.</p><p><strong>5. Governance and accountability: How is the work developed and represented?</strong></p><p>The quality of an idea and the way it’s governed are related, but they’re not interchangeable.</p><p>Institutional partnerships can provide resources, visibility and opportunities for collaboration. They can also introduce questions about independence, representation and conflicts of interest.</p><p>Those questions deserve examination.</p><p>Who contributes to the research? Are disabled people meaningfully involved where their experiences are relevant? How are disagreements handled? Are methods and limitations transparent? Can the work be challenged or revised?</p><p>These are more informative questions than whether a project has institutional backing.</p><p>A university, public body, charity or commercial organisation may support strong or weak work. Its identity alone doesn’t establish intellectual merit or improper intent.</p><p>The same principle applies to commercial activity.</p><p>Training, consultancy, publications and professional services are familiar parts of the accessibility sector. Their existence doesn’t automatically compromise research.</p><p>Actual conflicts of interest, exaggerated claims and misleading commercial practices are legitimate concerns. They should be examined through evidence about financial relationships, decisions and representations.</p><p>Speculation about what an emerging framework might eventually be used to sell tells us little about its present intellectual value.</p><p>Good governance requires transparency and accountability. It doesn’t require treating institutional visibility or possible future commercial activity as proof of wrongdoing.</p><p><strong>The same standards should apply to professional criticism</strong></p><p>These five criteria provide a way to examine emerging work without assuming in advance that it’s valuable or unnecessary.</p><p>They also raise a question about how criticism itself is evaluated.</p><p>Professional expertise is indispensable in accessibility. Experienced practitioners can identify overlooked research, challenge weak assumptions and explain practical consequences that may be missed by people approaching the field from another discipline.</p><p>But expertise should improve the quality of scrutiny, not replace it.</p><p>A criticism that identifies prior research should explain whether the new contribution genuinely duplicates it.</p><p>A concern about terminology should demonstrate where the language creates confusion.</p><p>A claim about institutional influence should identify the arrangements or consequences that justify concern.</p><p>And an allegation of harm should distinguish a plausible risk from an established outcome.</p><p>These aren’t special standards imposed on critics. They’re extensions of the same principles used to assess research.</p><p>There is a particular danger when professional familiarity begins to influence how those principles are applied.</p><p>Established terminology may be accepted without much examination, while new language is assumed to be unnecessary. Commercial activity may be treated as routine for recognised practitioners but suspicious for unfamiliar contributors. Institutional backing may signal credibility in one context and questionable motives in another.</p><p>Context can justify different judgements. But those differences should be explained rather than left to professional reputation.</p><p>A field that values evidence should be willing to examine the reasoning of its most experienced voices as carefully as it examines the work of newcomers.</p><p><strong>Intellectual openness isn’t the absence of standards</strong></p><p>Accessibility research doesn’t benefit from uncritical enthusiasm for every new framework.</p><p>Duplicative concepts, misleading terminology, weak evidence and exaggerated claims can all distract from important work.</p><p>They should be challenged.</p><p>But neither does accessibility benefit from treating established knowledge as though it has exhausted the possibilities for further inquiry.</p><p>The W3C’s continuing work on cognitive accessibility reflects an area that still requires research, practical development and attention to changing technologies and user needs.</p><p>Progress depends on preserving valuable existing knowledge while remaining open to better explanations and applications.</p><p>That requires intellectual discipline from everyone involved.</p><p>New frameworks should be clear about what they contribute, what remains untested and how their claims relate to previous scholarship.</p><p>Critics should be equally clear about the evidence supporting their objections and the limits of their conclusions.</p><p>Neither novelty nor reputation is a reliable substitute for careful analysis.</p><p>The question worth asking isn’t simply whether an idea is new, or whether a respected professional approves of it.</p><p>It’s whether the work contributes something that can be understood, examined and, where appropriate, tested.</p><p>That’s a standard capable of improving both research and the professional debate around it.</p><p><strong>Expertise may give an argument an audience. Evidence gives it authority.</strong></p><p><strong>Further reading</strong></p><ul><li>National Academies (2005), <em>Facilitating Interdisciplinary Research</em>: <a href="https://nap.nationalacademies.org/catalog/11153/facilitating-interdisciplinary-research">https://nap.nationalacademies.org/catalog/11153/facilitating-interdisciplinary-research</a></li><li>World Wide Web Consortium, <em>Cognitive Accessibility at W3C</em>: <a href="https://www.w3.org/WAI/cognitive/">https://www.w3.org/WAI/cognitive/</a></li><li>World Wide Web Consortium, <em>Making Content Usable for People with Cognitive and Learning Disabilities</em>: <a href="https://www.w3.org/TR/coga-usable/">https://www.w3.org/TR/coga-usable/</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=17b20c8eaec8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Neural Commons: Exploring the Emotional Dimensions of Digital Accessibility]]></title>
            <link>https://medium.com/@leenahaque1/neural-commons-exploring-the-emotional-dimensions-of-digital-accessibility-e7cebf40db75?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/e7cebf40db75</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[inclusive-design]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Mon, 12 Oct 2026 10:35:29 GMT</pubDate>
            <atom:updated>2026-10-12T10:35:29.717Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Editorial collage of diverse people, including a wheelchair user, alongside an AI figure and digital interface panels, with geometric shapes in blue, teal, yellow and coral." src="https://cdn-images-1.medium.com/max/1024/1*eMY8Xw_r8li0RWlyKIqYmg.png" /></figure><p><em>What the Neural Commons research tells us about emotional accessibility, digital trust and designing technology around human experience.</em></p><p><strong>Neural Commons</strong> is a collaborative initiative involving BBC CAPE, the University of Cambridge’s ThinkLab and Durham University’s Centre for Research into Inner Experience. It explores emotional accessibility: how the pace, tone, structure and behaviour of digital systems can influence people’s ability to engage with them.</p><p>I’ve been involved in developing Neural Commons, and one aspect of the work particularly interests me. When people describe a digital experience as exhausting, intimidating or difficult to trust, what can we learn about the design decisions that contributed to that experience?</p><p>It sounds like a straightforward question. In practice, it is difficult to answer.</p><p>Someone applying for financial assistance online may be able to navigate the service, understand the questions and submit their information. They may still be uncertain about whether an error can be corrected, what happens next or how to challenge a decision.</p><p>Some of that uncertainty comes from the circumstances. Financial insecurity is distressing regardless of how well a service is designed. But unclear explanations, unnecessary time pressure and inaccessible routes to support can make an already difficult situation harder.</p><p>I think this distinction matters. If we want to understand the emotional demands of digital technology, we need to examine what the system contributes to an experience without assuming that every uncomfortable feeling is a design failure.</p><p><strong>What Neural Commons is investigating</strong></p><p>The Neural Commons Collective’s August 2025 white paper, <a href="https://www.repository.cam.ac.uk/items/4dcbbdf2-c26c-4777-ba00-bb24e577dbf4"><em>The Right to Emotional Access in Digital Systems</em></a>, sets out a proposal for considering emotional experience more explicitly in digital inclusion.</p><p>The paper is grounded partly in an exploratory focus group involving 45 neurodivergent and neurotypical participants. Their feedback was organised around three themes: considerate interaction and pacing, digital trust, and inclusive design.</p><p>These are useful starting points because they concern recognisable features of everyday technology.</p><p>Participants described the pressure created by persistent notifications, the frustration of having to move through systems at a prescribed speed and the importance of being able to choose how they interact. They also raised concerns about manipulation, privacy and the possibility of technology monitoring their emotional state.</p><p>One of the clearest tensions in the findings is that people may want technology to be more responsive to their needs while also wanting it to make fewer assumptions about them.</p><p>That tension deserves attention as digital services become more personalised.</p><p>A system that offers a slower pace, fewer interruptions or an alternative way to communicate might improve someone’s experience. A system that decides how they are feeling and changes its behaviour without their knowledge could introduce a different problem.</p><p>The distinction is partly about control: who determines what kind of interaction is appropriate, and how easily can that decision be changed?</p><p><strong>Emotional accessibility and existing accessibility practice</strong></p><p>Digital accessibility has advanced through decades of standards development, research, advocacy and practical work led by disabled people and the wider accessibility community.</p><p>The Web Content Accessibility Guidelines and related W3C work provide essential foundations. Research into cognitive accessibility, usability and human–computer interaction already addresses matters such as predictability, comprehension, user control and the experience of interacting with technology.</p><p>Neural Commons builds on that landscape. It doesn’t require us to pretend that emotion is a new discovery in digital design, or that established accessibility practice has ignored people’s experiences.</p><p>The particular opportunity is to investigate the <em>relationship </em>between emotional experience and participation more deliberately.</p><p>Take the financial assistance application again. A friendly message saying “We’re here to help” might be welcome. But if the applicant cannot work out how to amend an answer, a reassuring tone will not resolve the underlying uncertainty.</p><p>A clearer explanation of the process, a visible route to human support or an opportunity to pause and return later may have greater practical value.</p><p>These are familiar design concerns. Emotional accessibility gives us another reason to examine their consequences and another set of questions to bring into research.</p><p>Did the person understand their options? Did the interaction create unnecessary pressure? Were they able to question what the system told them? Did the design support the level of independence they wanted?</p><p>The answers will depend on the person and the circumstances. An interaction that feels reassuring to one person may seem intrusive or patronising to another.</p><p>That variability is a reason for better research, not for a universal formula for how technology should make people feel.</p><p><strong>From principles to evidence</strong></p><p>The Neural Commons white paper proposes six design commitments: emotional legibility, cognitive respect, trustworthy atmosphere, rhythmic flexibility, nonlinear entry and affect awareness.</p><p>These are proposals for exploration rather than an established set of accessibility standards.</p><p>I find the commitment to rhythmic flexibility especially interesting because it raises questions that can be investigated through specific design choices.</p><p>Consider a service that requires someone to complete a sequence of steps within a limited period. What happens when the person is allowed to save their progress, take a break and resume without penalty? Does that change their understanding of the task, their willingness to continue or their sense of control?</p><p>A different experiment might examine how people respond to notifications. Would the ability to choose when reminders appear make the service easier to live with? What happens when the system suggests a break that the person neither needs nor wants?</p><p>The research challenge is to avoid treating a preferred feature as proof of an emotional benefit.</p><p>Someone completing a task more quickly may have found the experience easier, or may simply have wanted to leave. A participant reporting reassurance may have understood the process better, or may have been reassured by an explanation that was incomplete.</p><p>Good evaluation needs to account for those possibilities.</p><p>That means combining observations of interaction with people’s own descriptions of their experiences, examining what they understood and being careful about the conclusions drawn.</p><p>Durham University’s Centre for Research into Inner Experience brings relevant expertise to this challenge. Its wider research includes methods for investigating subjective experience, including Descriptive Experience Sampling. That does not mean any particular method has already been validated for emotional accessibility; it illustrates the kind of methodological expertise that makes the collaboration valuable.</p><p><strong>AI makes the question more complicated</strong></p><p>Adaptive AI systems can alter the experience of a digital service while someone is using it.</p><p>An assistant might change its tone, shorten an explanation or offer information in a different format. These adaptations could be useful, particularly when people can express their preferences directly.</p><p>But an AI system may also infer preferences incorrectly.</p><p>Imagine someone who usually asks for concise answers. When they later need to understand an important financial decision, the assistant continues to summarise. The response is readable but leaves out detail the person needs to assess their options.</p><p>They ask for more information. The assistant responds briefly again.</p><p>The problem is no longer just the original misunderstanding. It is the work the person must do to correct the system’s assumptions.</p><p>Research into personalisation, contestability and human–AI interaction already provides ways to examine such problems. Emotional accessibility brings attention to how that repeated negotiation is experienced.</p><p>Does the person feel able to challenge the assistant? Do they know whether it has changed its approach? Are they confident that important information has not been omitted?</p><p>A useful study could compare an interface where people choose their communication preferences with one where preferences are inferred. Researchers could then examine what happens when participants deliberately change their needs or correct the system.</p><p>There is a further ethical question here. The Neural Commons focus-group feedback included explicit concern about emotional monitoring. A system should not need to claim knowledge of someone’s inner state to offer them meaningful choices.</p><p>Giving people control over pacing, detail and interaction style may be a more appropriate starting point than attempting to detect how they feel.</p><p><strong>What I hope Neural Commons contributes</strong></p><p>The white paper sets out an ambitious direction, including proposals for research, design practice and possible future benchmarks. Turning those proposals into reliable evidence will take sustained work.</p><p>For me, the most promising contribution is the opportunity to bring accessibility expertise, design research and the study of inner experience into closer conversation.</p><p>The focus-group findings give that conversation a concrete basis. Participants have already identified concerns about pacing, manipulation, trust and being expected to adapt to systems built around narrow assumptions.</p><p>The next challenge is to understand those experiences with enough precision to inform design decisions.</p><p>That requires involving people with different lived experiences throughout the research. It also requires recognising that disabled and neurodivergent people are not uniform groups with identical emotional needs.</p><p>A design that offers choice must make those choices understandable and practical. A feature intended to reduce pressure should be tested to establish whether it does. And a system that appears empathetic should not be assumed to understand the person using it.</p><p>I would like Neural Commons to help develop evidence that designers and researchers can examine, challenge and apply. That is a more useful outcome than simply adding another appealing term to the language of digital design.</p><p>As AI becomes more embedded in everyday services, these questions will become harder to avoid. Systems will increasingly interpret what people want, decide how to respond and shape the conditions in which decisions are made.</p><p>We need to understand the consequences of those choices, particularly when people are dealing with uncertainty or important personal circumstances.</p><p>The question I keep returning to is this: <strong>when someone feels unable to continue, question or trust a digital interaction, how do we establish what role the design has played?</strong></p><p>Neural Commons offers a way to investigate that question collaboratively, drawing on existing knowledge while testing ideas that are still developing.</p><p>I think that is where its value can be demonstrated.</p><p><em>Author’s note: I have been involved in the development of BBC Neural Commons. This article offers my personal interpretation of the initiative’s research questions and published proposals. The views expressed are my own and do not necessarily represent those of the collaborating institutions.</em></p><p><strong>Sources and further reading</strong></p><ul><li><a href="https://api.repository.cam.ac.uk/server/api/core/bitstreams/96130db0-b3e7-4d09-8205-773c444f5668/content">Neural Commons Collective, <em>The Right to Emotional Access in Digital Systems</em> (August 2025)</a></li><li><a href="https://duracuk-lb01-production.terminalfour.net/research/institutes-and-centres/centre-for-research-into-inner-experience/projects/">Durham University, Centre for Research into Inner Experience — Projects</a></li><li><a href="https://www.w3.org/TR/coga-usable/">W3C Web Accessibility Initiative, <em>Making Content Usable for People with Cognitive and Learning Disabilities</em></a></li><li><a href="https://doi.org/10.3389/fpsyg.2017.00043">Hurlburt, Alderson-Day, Fernyhough and Kühn (2017), <em>Can inner experience be apprehended in high fidelity?</em></a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e7cebf40db75" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Best AI in Games Might Know When to Leave You Alone]]></title>
            <link>https://medium.com/@leenahaque1/the-best-ai-in-games-might-know-when-to-leave-you-alone-1f6f6d1b15d6?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/1f6f6d1b15d6</guid>
            <category><![CDATA[gaming]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Wed, 07 Oct 2026 11:12:54 GMT</pubDate>
            <atom:updated>2026-10-08T07:23:08.836Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Colourful collage of a gamer facing a difficult boss battle while a friendly AI robot reaches into the game interface and adjusts difficulty settings, illustrating the tension between helpful adaptation and player control." src="https://cdn-images-1.medium.com/max/856/1*6ezSLJ7kjCbHiWokyp46sg.png" /></figure><p><strong>AI is getting better at understanding players. That doesn’t necessarily mean it should act on everything it learns.</strong></p><p>There is something faintly insulting about discovering that a game has been going easy on you. Not when you’ve asked it to. Difficulty settings are useful, accessibility settings are essential, and there are plenty of games I have cheerfully switched to Easy because I had neither the time nor the desire to spend a Saturday afternoon being murdered by the same knight. I mean the other kind: the game quietly deciding that you have suffered enough.</p><p>You’ve been fighting the same boss for thirty minutes. You’re getting closer. You’ve learnt the first attack pattern, stopped making one stupid mistake and started making a more sophisticated stupid mistake instead. Then, eventually, you win. Except perhaps the game helped. Maybe it reduced the enemy’s aggression, gave you a little longer to react, or adjusted something so subtly that you never noticed. I think I’d care.</p><p>That question has been rattling around in my head while looking at what is happening with AI in games. We tend to talk about these systems in terms of capability: can the AI understand what is happening on screen, talk convincingly as a character, remember what the player has done, provide advice at the right moment? Increasingly, the answer is yes. What interests me more is what happens after the system understands you.</p><p><strong>Being stuck is sometimes the game</strong></p><p>Games have been adapting to players for years, long before anyone started attaching “AI” to every second product announcement. Dynamic difficulty adjustment is the obvious example. Instead of asking every player to experience exactly the same level of challenge, the game responds to how well you are doing.</p><p>There are perfectly sensible reasons for this. A game that is hopelessly difficult can become tedious. So can one that never asks anything of you. Designers have spent years trying to find the rather awkward space between boredom and throwing the controller at a wall.</p><p>There is research suggesting that adaptive difficulty can improve player experience, but it comes with an interesting complication. A <a href="https://doi.org/10.1145/3116595.3116623">CHI PLAY study from 2017</a>⁠ found benefits from dynamic difficulty systems, while also finding that one automated approach produced a lower sense of control. That part interests me more than the improvement score, because a system can make an experience smoother while still making the person using it feel that something has been taken away.</p><p>This is not a problem unique to games, but games make it unusually easy to see because friction is often part of what you came for. Most of UX has a fairly hostile relationship with friction. Too many clicks, remove them. Too much typing, autofill it. Too much thinking, recommend something. Too much waiting, make it instant. Generally, good. I don’t want my banking app to make checking my balance more character-building.</p><p>The logic starts to wobble when the effort itself has value. Anyone who has spent hours learning a game already knows this. The failed attempts are not necessarily unfortunate things happening around the experience. They can be the experience. Getting lost, misunderstanding something, trying again and eventually becoming competent is part of why the eventual success means anything.</p><p>So if an AI notices that I am struggling, it has learnt something useful about me. It hasn’t yet learnt whether I want help.</p><p><strong>The game is beginning to watch back</strong></p><p>This distinction matters more because the systems entering games are becoming much more capable of responding to individual players.</p><p>Microsoft has been developing <a href="https://news.xbox.com/en-us/2025/03/13/new-copilot-for-gaming-save-time-help-get-good/">Gaming Copilot</a>⁠, an assistant that can understand what is happening in a game and provide information or coaching. Microsoft’s own description is revealing: the idea is for it to be available when you need it and stay out of the way when you don’t. That sounds obvious until you ask who gets to decide when you need it.</p><p>NVIDIA’s work on <a href="https://www.nvidia.com/en-gb/geforce/news/nvidia-ace-autonomous-ai-companions-pubg-naraka-bladepoint/">ACE</a>⁠ is exploring AI characters that can perceive what is happening, reason about it and act inside games. Ubisoft’s NEO NPC experiments have been looking at characters capable of conversations that are considerably less scripted than the traditional choose-one-of-four-responses dialogue tree.</p><p>Much of this is genuinely exciting. Game characters have traditionally suffered from a peculiar kind of amnesia. You can save their village, insult their family, steal every object in their house and return ten minutes later to receive exactly the same line about the weather. Characters that remember and respond could make game worlds feel much less mechanical. An assistant that can explain an obscure system without forcing me to pause the game and find a seven-year-old Reddit thread sounds excellent. So does a companion that understands what I’m trying to do rather than walking repeatedly into a wall while I’m being shot at.</p><p>But as these systems get better at modelling the player, the design problem changes. It’s no longer simply about whether the game can react to me. It’s about how much authority I have given it to react.</p><p>You can imagine this becoming quite intimate. Games already collect enormous amounts of behavioural information. A sufficiently capable system does not need me to announce “I am frustrated now” in order to make a decent guess. Repeated failure, increasingly erratic inputs, abandoning a task, changing strategies or returning to the same section over and over are all signals.</p><p>There is also a growing body of research into affective games and systems that infer emotional state from behavioural or physiological information. The science here is not magically solved; emotion recognition remains messy, contextual and easy to overstate. But we don’t need perfect mind-reading for the design question to arrive. A system merely has to become good enough at guessing, and then somebody has to decide what it is allowed to do with that guess.</p><p><strong>Emotional accessibility is not emotional smoothing</strong></p><p>This is where the problem connects with my work on <a href="https://www.bbc.co.uk/neurodiversity/neural-commons/">Neural Commons</a>.</p><p>One of the ideas we have been exploring is emotional accessibility: what it would mean for digital systems to take the emotional and cognitive conditions of the person using them seriously. That can sound, at first, like an argument for making digital experiences gentler. I don’t think it is.</p><p>An emotionally accessible system shouldn’t have to make sure nobody is ever frustrated, frightened, uncertain, challenged or upset. Apart from being impossible, that would produce some extremely boring culture. What matters is the relationship between the person and the system. Do I understand what it is doing? Can I make meaningful choices? Does it demand attention it hasn’t earned? Can I change the terms of the interaction? Can I say no?</p><p>There is already related work happening specifically in games. Francesca Foffano and Paul Cairns’ CHI PLAY 2025 research, <a href="https://eprints.whiterose.ac.uk/id/eprint/229279/"><em>Do Not Play With My Emotions: Design for Emotional Accessibility</em></a>⁠, looks at how players encounter emotionally challenging material and how game design might better accommodate different needs. Their work approaches emotional accessibility from within games themselves. AI introduces another layer to the problem: what happens when a game is not simply presenting an emotional experience, but making judgements about mine?</p><p>Suppose the game correctly concludes that I am frustrated. It could offer a hint, lower the difficulty, change an enemy, point me towards a solution, have a companion character intervene or alter the pacing. Any of those things might genuinely improve my experience. They might also ruin it.</p><p><strong>Personalisation isn’t respectful merely because it’s accurate.</strong></p><p>That is the part of personalised UX I think we talk about far too little. We tend to assume that better knowledge of the user naturally leads to better service for the user. If Spotify knows my taste more accurately, recommendations improve. If a map knows where I am, directions improve. If an AI understands my context, assistance improves. But understanding creates power as well as usefulness, and the fact that a system has correctly inferred something about me doesn’t automatically give it permission to act on that information.</p><p>Games make this wonderfully obvious. If I choose Easy Mode, I’ve authored the adaptation. If I ask for a hint, I’ve authored the intervention. If the game secretly concludes that I’m having a rough evening and changes the rules until I win, the outcome may look identical on a dashboard. I progressed. I stayed engaged. Perhaps I even played for longer. But something different has happened to the experience.</p><p><strong>Perhaps AI should ask more often</strong></p><p>I don’t think the answer is to stop games adapting. For some players, adaptation can make the difference between being able to play and not being able to play. Difficulty, reaction time, cognitive load, sensory intensity and motor demands are not abstract design questions. Giving people more control over them is a good thing.</p><p>The interesting distinction is between adaptation and authorship. A genuinely useful AI might notice that I have failed the same section ten times and say, in effect: you seem to be stuck. Do you want a hint? Do you want the mechanic explained? Do you want this section made easier? Should I ask again later, or leave you alone?</p><p>That last option matters just as much as the others, because it is tempting to imagine increasingly intelligent systems becoming increasingly proactive. If the machine already knows what I probably need, why make me ask?</p><p>Because sometimes asking is the point.</p><p>The difference between assistance being available and assistance being imposed looks small on a product diagram. From the person’s side of the screen, it can be enormous. Think about an AI companion blurting out the solution to a puzzle three seconds before you solve it yourself, or a horror game deciding you’ve become sufficiently frightened and toning itself down when being frightened was the reason you bought it. Or an RPG character helpfully explaining a clue you were enjoying slowly piecing together.</p><p>All of these systems could be behaving intelligently. They could also be extremely annoying.</p><p>Humans are inconvenient users because our goals are not always reducible to comfort, speed or successful task completion. We voluntarily do difficult things. We choose suspense. We tolerate uncertainty. We replay sections we failed. Sometimes we spend an evening being repeatedly defeated by a fictional monster and consider the evening well spent.</p><p>Once you notice that, the gaming example starts leaking into the rest of AI. A writing assistant detects that you are hesitating and finishes the thought for you. Useful, unless the hesitation was the bit where you were deciding what you actually thought. A tutoring system detects confusion and provides the answer. Useful, unless working through the confusion was how the learning was going to happen. A fitness app detects discouragement and adjusts the target. A news product detects anxiety and starts quietly changing what it shows you. An AI companion learns which responses keep you talking and uses them more often.</p><p>None of those systems needs to be malicious. The system can be trying to help, the designer can be trying to help, and the prediction can even be correct. The intervention can still be wrong.</p><p>Games may turn out to be one of the better places to learn how to handle this because game designers already understand something the rest of technology occasionally forgets: the shortest route to an outcome is not necessarily the experience somebody wants. Sometimes I want the boss to stay difficult. Sometimes I want to wander around lost for twenty minutes. Sometimes I want to work it out myself.</p><p>And sometimes, after an AI has correctly worked out that I am struggling, the most useful thing it can do is leave me alone.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1f6f6d1b15d6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why We Trust News Creators More Than the News]]></title>
            <link>https://medium.com/@leenahaque1/why-we-trust-news-creators-more-than-the-news-7c7b253b3bf0?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/7c7b253b3bf0</guid>
            <category><![CDATA[creator-economy]]></category>
            <category><![CDATA[media]]></category>
            <category><![CDATA[journalism]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Wed, 07 Oct 2026 11:03:15 GMT</pubDate>
            <atom:updated>2026-10-07T11:03:15.876Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Mixed media style collage contrasting creator-led and traditional news: a woman speaks into a microphone on a smartphone surrounded by social media portraits, facing newspapers, a TV news anchor, satellite dish and classical “NEWS” building." src="https://cdn-images-1.medium.com/max/1024/1*D6vnD7xI3b3R-Wnrqh-UlQ.png" /></figure><p><strong>Trust in news isn’t disappearing. We may be taking it away from institutions and handing it to people we think we know.</strong></p><p>There is a particular kind of person who will tell you, with total confidence, that they no longer trust the media.</p><p>Fair enough. But then something happens in Ukraine, Westminster or Washington and they send you a 14-minute YouTube video made by a man speaking into a microphone from what appears to be his spare bedroom.</p><p>“You should watch this guy. He actually explains what’s going on.”</p><p>I’m fascinated by that sentence because this person hasn’t stopped trusting the news at all. They’ve changed who gets to deliver it.</p><p>For much of the last century, trust in news belonged to institutions. You bought a newspaper or turned on a broadcaster and, however imperfectly, the name above the door did some of the work. The BBC was the BBC. Reuters was Reuters. A masthead represented editors, correspondents, standards, lawyers, foreign bureaux and a large number of people whose names you would never know.</p><p>Open TikTok and the arrangement is rather different. There is a face looking back at you, and the face has an advantage the institution doesn’t.</p><p>You like him.</p><p><strong>“Watch this guy”</strong></p><p>Listen to the way people recommend news now. “The Financial Times has a good analysis of this” still exists, but increasingly so does “Watch this guy,” or “She’s really good on this.”</p><p>My favourite endorsement might be: “He’s actually pretty balanced.”</p><p>The interesting word isn’t <em>balanced</em>. It’s <em>he</em>.</p><p>The judgement has moved from the publication to the person.</p><p>It isn’t necessarily because that person has better sources, more experience or a team of editors checking every sentence. Often it’s because you’ve spent a surprising amount of time with them.</p><p>You know the room they record in. You know the mug. You know the slightly annoying thing they do when they’re excited. You know which subjects make them angry and which ones they admit they don’t understand.</p><p>Perhaps you’ve watched them get something wrong. More importantly, perhaps you’ve watched them say they got something wrong.</p><p>That counts for a lot.</p><p>After two years of this, you aren’t really consuming a media product in the old sense. You have a relationship with a person who happens to tell you about the world. It’s a strange relationship, admittedly. They probably don’t know you exist. Your brain doesn’t seem especially bothered by that detail.</p><p><strong>The world’s smallest trusted newsroom</strong></p><p>This produces a form of authority that would have looked bizarre twenty years ago.</p><p>A creator can be trusted intensely by 100,000 people and be completely unknown outside that group.</p><p>Try doing that with a newspaper.</p><p>You don’t have to read The Guardian to have an opinion about The Guardian. You don’t have to watch the BBC to know what the BBC represents. Institutional reputations travel beyond their audiences.</p><p>Creator reputations often don’t.</p><p>A political YouTuber with 300,000 subscribers might be regarded by those subscribers as one of the most sensible people covering British politics. Mention her at dinner and nobody knows who you’re talking about.</p><p>This isn’t celebrity in the old sense. It’s something narrower and, in some ways, more powerful.</p><p>An institution can be famous without being trusted. A person can be trusted without being famous.</p><p>Journalism is beginning to discover that the second category matters quite a lot.</p><p><em>[The Reuters Institute’s research into news creators provides useful evidence for this shift: more than a quarter of people surveyed use news creators weekly, while only a small minority rely on them for all their news. Link to the original research </em><a href="https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2026/dnr-executive-summary"><em>here</em></a><em>.]</em></p><p><strong>The personality is not a bug</strong></p><p>Traditional journalism spent a long time trying to remove the journalist from the journalism.</p><p>There were good reasons for this. The reporter wasn’t supposed to be the story. Facts were meant to survive independently of whether you liked the person presenting them.</p><p>Creators have more or less inverted the arrangement.</p><p>The person is part of the product.</p><p>They can admit they don’t understand something. They can spend twenty minutes wandering around one obscure point because they find it interesting. They can tell you what they think and then show you how they got there.</p><p>Sometimes this produces dreadful journalism. Sometimes it produces something oddly refreshing.</p><p>Audiences don’t necessarily experience an absence of personality as an absence of bias. Quite often they just experience it as corporate language.</p><p>“I have a view and here it is” gives you something to interrogate.</p><p>The polished institutional voice can be harder to read. Who decided this angle? Why this headline? Why this expert? Why isn’t that fact higher up?</p><p>The creator, meanwhile, is sitting there telling you exactly what annoyed her about the story.</p><p>You may disagree with her, but at least you know where she is standing.</p><p><strong>From attention to trust</strong></p><p>This is a question I’ve been thinking about through my work on Neural Commons.</p><p>Neural Commons started from a different problem: what would digital systems look like if they respected the emotional and cognitive lives of the people using them?</p><p>We use the term “emotional accessibility” for some of this work. In practice, that means asking questions technology has become surprisingly good at avoiding. Does this interaction deserve your attention? Does it need to interrupt you? Does it give you enough information to make a choice? Can you leave and come back without being punished for it?</p><p><em>[Link to the BBC Neural Commons work / white paper </em><a href="https://www.bbc.co.uk/neurodiversity/neural-commons/"><em>here</em></a><em>.]</em></p><p>Once you start asking those questions, the attention economy begins to look rather strange.</p><p>For years, we have treated attention as something digital products are entitled to compete for. The notification, the red badge, the breaking-news banner, the headline engineered to create just enough anxiety or curiosity to make you click.</p><p>News inherited the same logic.</p><p>But while working on Neural Commons, I’ve become increasingly interested in a different possibility: what if the scarce resource isn’t attention, but trust?</p><p>Or, more precisely, <strong>attention willingly given</strong>.</p><p>Watch somebody disappear into TikTok for 90 minutes and it’s difficult to maintain that human beings have run out of attention. We have plenty of it.</p><p>The harder question is who gets it without having to steal it.</p><p>A notification can make me look at my phone. It can’t make me want to hear from you tomorrow.</p><p>Trust can.</p><p><strong>Trust doesn’t replace attention. It changes the terms on which attention is given.</strong></p><p>And this is where news creators become particularly interesting.</p><p>When somebody returns to the same journalist, YouTuber or commentator every week, they’re doing something different from responding to a headline. They’re building up a history with that person.</p><p>They learn what the creator chooses to cover and what they ignore. They notice how uncertainty is handled. They remember predictions that didn’t come true. They see whether mistakes are acknowledged or quietly allowed to disappear down the timeline.</p><p>Over time, the audience builds a model of the person.</p><p>Traditional media organisations have spent extraordinary amounts of money trying to manufacture engagement. Creators, often with a microphone and a camera in a spare bedroom, have stumbled onto something potentially more valuable: return.</p><p>Not simply: <strong>Can I make you click?</strong></p><p>But: <strong>Will you come back when nobody reminds you?</strong></p><p>That’s the connection I keep coming back to between Neural Commons and journalism.</p><p>The internet became remarkably good at learning how to capture our attention. The next problem is learning how to deserve it.</p><p><strong>What would a trust economy actually reward?</strong></p><p>A trust economy wouldn’t mean abandoning attention. Journalism obviously needs to be noticed.</p><p>But it might change what we optimise for.</p><p>“We don’t know yet” could become a sign of credibility rather than weakness. Not sending another notification might occasionally be the respectful editorial decision. Being worth someone’s time could matter more than successfully stealing thirty seconds of it.</p><p>This sounds like a subtle distinction until you think about what digital publishing currently measures.</p><p>Clicks. Views. Watch time. Completion rates. Shares. Comments. Subscribers.</p><p>All useful. None tells you quite the same thing as: <strong>Would this person choose us again?</strong></p><p>Creators have stumbled into parts of this model because their relationship with an audience already depends on return. People aren’t simply following stories anymore. They’re following people.</p><p>That creates a harder problem for traditional news organisations.</p><p>They have spent decades asking audiences to trust the institution.</p><p>Increasingly, audiences may be asking which people inside that institution they trust.</p><p>There’s an opportunity in that. Newsrooms already contain journalists with expertise, judgement, humour, obsessions and recognisable ways of looking at the world. For a long time, institutional journalism treated much of that personality as something to smooth out.</p><p>Perhaps some of it should be put back in.</p><p><strong>There is, unfortunately, a catch</strong></p><p>All of this sounds lovely until you remember that humans are spectacularly capable of trusting the wrong person.</p><p>The boring machinery of institutional journalism exists for a reason.</p><p>A reporter has an editor. The editor asks irritating questions. Lawyers become involved when necessary. Claims are checked. Somebody asks whether the headline actually matches the story.</p><p>The system can fail, but at least there is a system to fail.</p><p>The independent creator may be reporter, editor, presenter, publisher, salesperson and chief executive.</p><p>Sometimes the fact-checking department is the comments section.</p><p>Audiences aren’t neutral either.</p><p>Imagine discovering that whenever you say X, 400,000 people applaud, your subscriber count rises and several thousand pounds appear in your bank account.</p><p>Then imagine discovering that whenever you say Y, those same people become annoyed and unsubscribe.</p><p>How many times do you have to run that experiment before your worldview starts mysteriously producing more X?</p><p>You don’t need to be dishonest. You barely need to notice it happening.</p><p>The audience trains the creator while the creator trains the audience.</p><p>That’s where relationship-based trust gets uncomfortable.</p><p><strong>The attention economy hasn’t gone anywhere</strong></p><p>Creators haven’t escaped the attention economy. They’re standing right in the middle of it.</p><p>They need views. Platforms need engagement. Algorithms discover that outrage works. Creators discover that outrage works. Everyone is shocked by the resulting outrage.</p><p>The difference is that creators possess something many institutions would desperately like to recover: a personal relationship with the audience.</p><p>And that relationship is valuable enough to exploit.</p><p>The person who spent three years earning your trust eventually faces the same temptation as the newspaper you stopped reading.</p><p>Attention pays. The uncomfortable possibility is that trust pays even better.</p><p><strong>Maybe the trust crisis isn’t a trust crisis</strong></p><p>This is why I’m increasingly unconvinced by the familiar story that people have simply “lost trust in news.”</p><p>Many clearly still trust people who tell them about the news. They just don’t necessarily trust the organisations those people used to work for.</p><p>Trust hasn’t vanished. It has become more personal, more fragmented and probably harder to measure.</p><p>Millions of people can distrust “the media” while maintaining intense confidence in their own little constellation of podcasters, Substack writers, TikTok explainers, YouTubers and journalists.</p><p>Everyone thinks everyone else is being manipulated.</p><p>Everyone has their guy.</p><p>Perhaps that is the media landscape we’re actually entering.</p><p>The important question isn’t whether institutional trust or creator trust wins. Neither deserves automatic loyalty. The better question is what our trust allows the person receiving it to get away with.</p><p>Someone worth trusting should be able to disappoint you occasionally. They should be able to tell you something you don’t want to hear. They should be able to say “I was wrong” without destroying the relationship.</p><p>And you should still be capable of thinking they’re wrong after watching them every morning for three years.</p><p>Because recognising someone’s face, knowing their politics and liking the way they explain the world tells you quite a lot about why you trust them.</p><p>It still doesn’t tell you whether they’re right.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7c7b253b3bf0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Seen by 23]]></title>
            <link>https://medium.com/@leenahaque1/seen-by-23-4346be0515c8?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/4346be0515c8</guid>
            <category><![CDATA[digital-culture]]></category>
            <category><![CDATA[south-asian-diaspora]]></category>
            <category><![CDATA[design]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Fri, 02 Oct 2026 20:58:05 GMT</pubDate>
            <atom:updated>2026-10-02T20:58:05.978Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Illustration of a South Asian family using WhatsApp across generations, centred on a family group chat showing a good-morning message, reactions and “Seen by 23,” representing digital kinship and diaspora family life." src="https://cdn-images-1.medium.com/max/1024/1*0nbojc8RRs7rUD4hgC6bog.png" /></figure><p>At 7:14 in the morning, a family WhatsApp group receives a rose. It is extremely red, sitting in front of a lake, two swans and a sunrise. Across the water are the words <strong>GOOD MORNING. HAVE A BLESSED DAY.</strong></p><p>Nobody replies. By nine, twenty-three people have seen it.</p><p>There isn’t much information in the message. Nobody has asked a question or shared any news. Tomorrow’s image may contain a different flower and essentially the same greeting.</p><p>Yet if the person who has sent one of these images every morning for three years suddenly stops, somebody will notice.</p><p>One relative finds the roses sentimental. Another barely looks at them. Someone is probably irritated by the amount of space they take up on their phone. For the person sending them, it may simply be a way of appearing in the family each morning without needing a reason to start a conversation.</p><p>Twenty-three people see the same rose. They don’t necessarily receive the same message.</p><p><strong>WhatsApp has no word for</strong></p><p><strong>phuppo</strong></p><p>South Asian families can have remarkably precise words for relationships. A <em>phuppo</em> is not a <em>khala</em>. A <em>mamu</em> is not a <em>chacha</em>. The words tell you where someone sits in the family, but years of affection, obligation, irritation and history tend to come attached.</p><p>Then the family enters WhatsApp and everyone becomes a <strong>participant</strong>.</p><p>A grandmother, an eldest brother, a cousin nobody has seen for eight years, a new daughter-in-law and the uncle sending the morning roses all sit in the same list.</p><p>It looks like a fairly brutal compression of family life.</p><p>Except people put the missing detail back.</p><p>A message from a <em>phuppo</em> is still a message from a <em>phuppo</em>, even if WhatsApp knows her as a name and phone number. A heart from a cousin you speak to every day feels different from one sent by somebody who rarely shows affection. Three missed calls from one relative might be ordinary; three from another and you call back immediately.</p><p>The app doesn’t need to explain these differences. People arrive already knowing many of them.</p><p>The relationship supplies the missing information.</p><p><strong>A surprisingly small vocabulary</strong></p><p>The basic materials of a family group are quite limited: messages, photographs, reactions, read receipts, forwards and a handful of controls.</p><p>Families manage to make a lot out of them.</p><p>Take ❤️.</p><p>A cousin posts a photograph of a new baby and eighteen hearts appear underneath. On the screen, eighteen people have done almost exactly the same thing.</p><p>In the family, they may have done eighteen slightly different things.</p><p>One heart is affectionate. Another is dutiful. One comes from somebody who speaks to the parents every day; another from a cousin who hasn’t seen them since childhood. Someone presses it because writing another “Congratulations!” feels unnecessary. Someone else presses it because saying nothing would feel noticeable.</p><p>The symbol itself carries very little. Most of its meaning comes from knowing who sent it.</p><p>This is <strong>social compression</strong>. A tiny signal travels through the app, and the people receiving it fill in what is missing from years of shared context.</p><p>Sometimes they get it wrong, of course.</p><p>The blue ticks say a message was opened. They don’t say why nobody replied. Perhaps the person was busy. Perhaps they forgot. Perhaps they were annoyed. Perhaps there is an argument happening that half the group knows about and the other half doesn’t.</p><p>The interface leaves that part with the family.</p><p><strong>Some messages barely contain messages</strong></p><p>This makes the good-morning rose more interesting than it first appears.</p><p>After the hundredth version, nobody needs reminding that mornings exist. But information may be the wrong way to judge it.</p><p>A parent asking an adult child whether they have eaten is not necessarily looking for a detailed account of lunch. A relative forwarding a prayer may be saying something that isn’t written in the prayer. A photograph from a family gathering can matter even when everyone already knows the gathering happened.</p><p>The message is small because the relationship is doing much of the work.</p><p>You could call this <strong>low-information, high-context communication</strong>.</p><p>It becomes especially useful when families are spread across countries. Keeping every relationship alive through long conversations would be exhausting, and probably quite strange. Most people don’t have enough to say to thirty relatives every week.</p><p>So smaller signals take over.</p><p>A heart appears under a photograph. Someone changes their profile picture. A relative who rarely posts suddenly congratulates somebody. A familiar name appears under “read by”.</p><p>None of these amounts to much on its own. Over time, they can stop people from disappearing from one another’s lives.</p><p><strong>Knowing someone without talking to them</strong></p><p>Most large family groups have quiet members.</p><p>They don’t send morning greetings. They rarely react. Months can pass without their name appearing in the conversation.</p><p>Yet they have watched children grow up, seen new houses, weddings, holidays and graduations. Someone might recognise a cousin’s kitchen in another country despite never having visited it.</p><p>Calling this “keeping in touch” feels slightly wrong. Very little direct contact may be happening.</p><p>They are simply present at the edge of one another’s attention.</p><p><strong>Ambient kinship</strong> might be a useful name for it: family relationships maintained partly through signals too small to become conversations.</p><p>That doesn’t necessarily mean people are close. Knowing what someone’s children look like is not the same as knowing how they’re doing. Sometimes digital familiarity probably disguises how little two people actually know about one another.</p><p>But migration has always left families with the problem of what happens to relationships when ordinary life is no longer shared.</p><p>Messaging apps have made one particular thing very easy: remaining visible.</p><p>A cousin doesn’t have to become a stranger just because nobody has found a reason to call. They can hover pleasantly in the background for years.</p><p><strong>The photograph where Nana was outside the old house</strong></p><p>Someone posts an old family photograph and six relatives begin trying to identify everyone in it. One remembers the year. Another says they’re wrong. Someone knows the street. A cousin who hasn’t written in the group for months appears because they recognise the man standing at the back.</p><p>The file itself holds very little of this.</p><p>A family might remember it simply as <em>the photograph where Nana was outside the old house</em>.</p><p>That description would be useless to almost anyone else. Inside the family, it is surprisingly precise. “Nana” assumes everyone knows whose Nana. “The old house” assumes everyone knows which house became old enough to earn the name. Someone remembers the front step. Somebody else remembers who lived next door.</p><p>The photograph isn’t carrying all that information. The people looking at it are.</p><p><strong>Maybe the missing information is useful</strong></p><p>There is an obvious temptation here for design.</p><p>If software flattens all these relationships into “participants”, perhaps the answer is to make it understand families better. Give relationships labels. Build family structures into profiles. Make the interface more aware of who is close to whom.</p><p>But would that actually improve things?</p><p>Imagine WhatsApp knowing not only that someone is an aunt, but which kind of aunt. Imagine it knowing which relatives are close, which relationships have cooled, who usually replies immediately and who has recently stopped.</p><p>At some point, better understanding starts to feel less like thoughtful design and more like an uncomfortable amount for an app to know about a family.</p><p>Perhaps the simplicity is doing useful work.</p><p>WhatsApp doesn’t have to know what a <em>phuppo</em> is because the family already does.</p><p>The same blank category, <strong>participant</strong>, can hold a grandmother, an estranged cousin, a childhood friend treated as family or an aunt whose exact relationship takes five minutes to explain. The software leaves the details alone, and people carry them into the space themselves.</p><p>There may be an advantage in that kind of simplicity. The interface doesn’t need to describe the family before the family can use it.</p><p>People fill the gaps.</p><p>That doesn’t always go smoothly. A heart can be affectionate or dutiful. Silence can be comfortable or pointed. The blue ticks tell you someone saw the message, then leave everyone to work out what happened next.</p><p>But perhaps that unfinished quality matters.</p><p>Twenty-three people see the same rose. The app counts twenty-three views.</p><p>One person smiles. Another ignores it. Someone finds it embarrassing. Someone else would worry if tomorrow morning it didn’t arrive.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4346be0515c8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Emotional Accessibility Without Emotion Detection: Designing AI That Gives People Control]]></title>
            <link>https://medium.com/@leenahaque1/emotional-accessibility-without-emotion-detection-designing-ai-that-gives-people-control-0e26293cab15?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/0e26293cab15</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[human-centered-design]]></category>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[neurodiversity]]></category>
            <category><![CDATA[ux-design]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Wed, 30 Sep 2026 13:31:02 GMT</pubDate>
            <atom:updated>2026-09-30T13:31:02.092Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3jdoKuY0Ig7rz4MUwOlI3w.png" /></figure><p>There’s an assumption hiding inside the idea of emotional accessibility: if a digital system is going to respond to how people feel, it must first work out what they are feeling.</p><p>But emotional accessibility and emotion detection are not the same thing. A service can give people more time, control and predictability without trying to decide whether they are anxious, frustrated or overwhelmed.</p><p>This distinction matters more as AI becomes part of everyday interfaces. An ordinary website might display the wrong information or make a process unnecessarily difficult. An AI assistant can go further. It can interpret what somebody means, adapt its responses, make recommendations and, increasingly, take actions on their behalf. If those decisions are based on assumptions about the person using it, a mistaken assumption can begin to shape the experience itself.</p><p>Emotion detection takes several forms. Systems can analyse facial expressions, speech, text, behaviour and other signals in an attempt to infer emotional states. Context is a significant challenge in this field: tone of voice, body language, facial expression and the situation itself can all affect how an apparent emotional signal is interpreted.</p><p>There is another way to approach the problem.</p><p>Imagine you are halfway through a long online application when a message appears telling you that the session is about to expire. You might have been checking information in another document. You might need longer to read the questions. You might simply have been interrupted.</p><p>The service does not need to know which explanation is correct. It could save your work, warn you in good time and provide a straightforward way to extend the session.</p><p>No emotional inference is required.</p><h4>Start with what the system is doing</h4><p>A surprising amount of our experience of technology is shaped by fairly mundane product decisions.</p><p>How long does someone have to complete something? What happens when they leave and return? How often does a service interrupt them? Does a dismissed message stay dismissed? Can an unexpected change be reversed? What happens when an automated action is wrong?</p><p>Those are design questions before they are questions about emotion.</p><p>Consider notifications. A product team doesn’t need to establish whether notifications make a particular person anxious before giving them better control over interruptions. People can be allowed to group notifications, silence them temporarily or decide which ones justify an immediate interruption.</p><p>The same applies to an AI assistant that keeps making unwanted suggestions. The problem may not be that the assistant has failed to recognise irritation. It may simply lack a dependable way for somebody to say, “I don’t want this suggestion again.”</p><p>This is close to an established principle in cognitive accessibility: people can benefit from being able to adapt and personalise an interface, including controlling content that moves or changes. <a href="https://www.w3.org/WAI/WCAG2/supplemental/objectives/o8-personalization/">W3C’s supplemental cognitive accessibility guidance </a>explicitly encourages support for adaptation and personalisation.</p><p>Emotional accessibility adds another question: <strong>what conditions does the behaviour of the system create for the person using it?</strong></p><h4>Emotion detection is a different proposition</h4><p>There are legitimate reasons to research how technology responds to human emotion. But doing so introduces a different set of technical and ethical problems.</p><p>Human behaviour is highly dependent on context. Someone who pauses for a long time might be confused, but they might equally be thinking carefully, reading something elsewhere or dealing with an interruption. Short answers do not necessarily indicate frustration. A particular facial expression or tone of voice does not have one reliable meaning in every person and every situation.</p><p>This isn’t merely a hypothetical concern. <a href="https://www.sciencedirect.com/science/article/abs/pii/S0925231224018447">Research into context-based emotion recognition</a> treats situational cues, body language, tone and other contextual information as important to interpretation. The <a href="https://ai-act-service-desk.ec.europa.eu/en/ai-act/recital-44">EU AI Act</a> goes further in its reasoning around biometric emotion inference, citing concerns about limited reliability, specificity and generalisability, as well as variation across cultures, situations and individuals. The Act prohibits certain uses of AI emotion inference in workplaces and educational institutions, subject to limited medical and safety exceptions.</p><p>Neurodiversity makes simple assumptions particularly problematic. Expressions of discomfort, concentration, excitement or overload can vary substantially between people, and systems developed around typical patterns of expression may not generalise reliably to everyone.</p><p>A <a href="https://www.sciencedirect.com/science/article/pii/S1386505624001321">2024 systematic review</a> of emotion-recognition technologies in autism found practical challenges in applying general emotion-recognition systems, noting that these systems are normally developed for and by neurotypical people. The review also found limited validation in autism and relatively little attention to privacy and security in the literature it examined.</p><p>Imagine, then, an AI assistant deciding that somebody appears confused and automatically simplifying its responses. That could be useful. It could also be irritating if the person understood the subject perfectly well and simply communicated in a way the model interpreted incorrectly.</p><p>At that point, accuracy is only part of the design problem. We also need to know whether the person can see that the system has adapted, change the adaptation and prevent an incorrect assumption from quietly following them into future interactions.</p><h4>Before improving an inference, ask whether it is necessary</h4><p>AI design places considerable emphasis on prediction. If a system learns enough about somebody, the argument goes, perhaps it can anticipate what they need before they have to ask.</p><p>Sometimes that will be useful. But prediction is not the only way for an interface to be responsive.</p><p>People can tell systems what works.</p><p>Someone might want an assistant to be proactive while planning a journey but quiet while they are writing. They might prefer short explanations during an unfamiliar administrative process and detailed ones when discussing a subject they know well. The reminders they appreciate on Monday might be exactly the interruptions they want to avoid on Friday.</p><p>A preference is not necessarily a characteristic of a person. Sometimes it is simply a preference for this task, in this context, at this moment.</p><p>That distinction matters because a system built around persistent inference can turn temporary behaviour into an enduring assumption. Something that was useful yesterday can become a nuisance today.</p><p>Controls such as “give me more time”, “ask before doing that again” and “stop suggesting this” are not technologically impressive. Their value is that they allow a person to communicate a need without having to demonstrate, disclose or perform an emotional state first.</p><p>That suggests a useful rule for emotionally accessible AI:</p><p><strong>Before trying to improve an inference, ask whether the inference is necessary.</strong></p><h4>AI makes control more consequential</h4><p>The issue becomes more important as software moves beyond presenting information.</p><p>AI can already reorganise complicated material, change the level of detail in an explanation, recommend what somebody should do next and assist with unfamiliar processes. Agentic systems extend this further by taking actions rather than merely generating responses.</p><p>These capabilities could remove genuine barriers. They also create the possibility of a system becoming easier to operate while becoming harder to control.</p><p>Suppose an assistant notices that somebody regularly requests concise answers and begins shortening responses automatically. At first, this looks like successful personalisation. Later, the person asks about something complicated and receives a summary when they wanted detail.</p><p>Better emotion detection does not necessarily solve that problem. What would help is knowing that the response has been adapted, being able to change it easily and not having a temporary preference become difficult to escape.</p><p>The same principle becomes more important as the consequences increase. Correcting the length of an answer is relatively trivial. Correcting an AI after it has cancelled something, submitted something or communicated with another person may not be.</p><p>Emotional accessibility therefore starts to intersect with agency, reversibility and meaningful control.</p><h4>Personalisation needs an exit</h4><p>Personalisation is often presented as an uncomplicated improvement: the more accurately technology understands us, the better it can serve us.</p><p>In practice, another question matters just as much: how much control does somebody retain over what the system thinks it knows?</p><p>Yesterday’s useful inference can become today’s nuisance. An AI might conclude from previous interactions that somebody wants simplified information, fewer choices, a particular tone or frequent assistance. Even where the original inference was reasonable, circumstances can change.</p><p>Where practical, inferred preferences should therefore be visible enough to be contested. People should be able to correct them, reset them or make them temporary rather than permanent.</p><p>There should also be room for a less fashionable possibility: somebody may not want to be personalised at all.</p><p>The aim shouldn’t necessarily be an AI that knows as much as possible about the person using it. In some circumstances, an AI that knows less but responds reliably to explicit choices may provide the more accessible experience.</p><h4>What this means for emotional accessibility</h4><p>This is one of the questions we are exploring through <a href="https://www.bbc.co.uk/neurodiversity/neural-commons/">Neural Commons</a>.</p><p>Emotional accessibility is concerned with the conditions people encounter when they use digital systems, including pressure, interruption, predictability, adaptation and control. There is a clear relationship with cognitive accessibility and neuroinclusive design, particularly where the behaviour of an interface creates additional demands for the person using it.</p><p>Taking emotional experience seriously, however, doesn’t automatically mean collecting more information about emotion.</p><p>There is another research direction available: examining whether systems can become more accommodating while requiring fewer assumptions about the internal state of the person using them.</p><p>For AI, that could mean investigating whether adaptations should be visible, whether inferred preferences should expire, how easily unwanted personalisation can be reversed, whether a system respects a request not to repeat an intervention, and when explicit choice works better than automated inference.</p><p>These are testable questions. We can observe whether people understand the controls available to them. We can test whether they can recover from unwanted adaptations. We can examine whether an AI continues to apply an assumption after it has been corrected.</p><p>That gives emotional accessibility something concrete to investigate rather than requiring designers to establish what somebody <em>should</em> be feeling while they use a system.</p><h4>Sometimes good design is enough</h4><p>There will be applications where people actively choose to share information about their emotional state and find value in technology responding to it. Nothing about emotional accessibility requires those possibilities to be rejected.</p><p>The more useful question is whether emotional inference is necessary for the problem at hand.</p><p>Someone who needs longer to complete a process can be offered more time. Interruptions can be controlled. An unwanted AI adaptation can be corrected. Personalisation can have an off switch. Consequential automated actions can include opportunities for confirmation and, where possible, reversal.</p><p>Most of this is ordinary interaction design. AI doesn’t make those principles obsolete. It makes them more important because the system has greater capacity to interpret, adapt and act.</p><p>A practical starting point for emotionally accessible AI may therefore be quite modest: <strong>minimise unnecessary inference, make adaptation visible, allow preferences to change, and preserve meaningful ways for people to intervene.</strong></p><p>The goal is not to build a machine that always knows how somebody feels. It is to build digital systems that don’t need to know in order to give people room to decide for themselves.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0e26293cab15" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Emotional Accessibility: Designing the Conditions of Digital Access]]></title>
            <link>https://medium.com/@leenahaque1/emotional-accessibility-designing-the-conditions-of-digital-access-f9c81f9dfa70?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/f9c81f9dfa70</guid>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[neurodiversity]]></category>
            <category><![CDATA[design]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Tue, 29 Sep 2026 15:01:02 GMT</pubDate>
            <atom:updated>2026-09-29T15:01:02.454Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uSZ44E1O6HK8P9pG1zQPHg.png" /></figure><p><em>How Neural Commons frames pace, pressure, tone and trust as conditions of digital access.</em></p><p>A digital service can be readable, understandable and technically accessible yet still make someone feel rushed, pressured or reluctant to continue. That gap is where emotional accessibility becomes useful.</p><p>Accessibility practice has developed sophisticated ways of identifying visual, auditory, physical and cognitive barriers. What receives less systematic attention is the emotional character of digital interaction itself: the way tone, pace, pressure and behavioural design affect a person’s ability to participate.</p><p>This matters because digital systems do more than provide information. They establish a rhythm, decide when to interrupt, determine how easily someone can pause, frame choices through language and increasingly attempt to shape behaviour. Those decisions form part of the environment in which access takes place. Emotional accessibility offers a way of examining that environment more directly.</p><p><strong>What emotional accessibility means</strong></p><p>In 2025, the <a href="https://www.bbc.co.uk/neurodiversity/neural-commons/">Neural Commons Collective</a> — involving <a href="https://www.bbc.co.uk/neurodiversity">BBC CAPE</a>, the <a href="https://www.thinklab.strategic-partnerships.admin.cam.ac.uk/">University of Cambridge ThinkLab</a> and <a href="https://www.durham.ac.uk/research/institutes-and-centres/centre-for-research-into-inner-experience/">Durham University’s Centre for Research into Inner Experience</a> published the white paper <a href="https://www.bbc.co.uk/neurodiversity/neural-commons-white-paper/"><em>The Right to Emotional Access in Digital Systems</em></a>. It proposes emotional access as an additional dimension of digital inclusion, positioned alongside existing accessibility and rights frameworks.</p><p>The work is exploratory, with its language and methods still developing. Its brings together concerns around emotion, stress, wellbeing and human-centred interaction together and treat the emotional conditions of interaction as an explicit accessibility concern.</p><p>The paper is concerned with whether people can use a system in a way that is emotionally safe, cognitively manageable and compatible with how they make sense of the world. It pays particular attention to pace, emotional tone, trust, sensory and affective forms of interaction, and the ability to engage without being forced into a single preferred rhythm.</p><p>Consider a cancellation journey that is easy to read and simple to navigate while using guilt to discourage somebody from leaving. A public service can meet technical accessibility requirements while its language feels threatening or accusatory. A page can be straightforward to understand while an unnecessary countdown introduces pressure. An otherwise usable process can become difficult to tolerate through repeated prompts or an inability to pause without losing progress.</p><p>In each case, the user may know exactly what the system is asking them to do. The difficulty does not necessarily lie in comprehension or operability. It lies in the conditions under which the interaction is taking place. Thinking in terms of emotional accessibility makes those conditions easier to examine directly.</p><p><strong>Cognitive and emotional accessibility</strong></p><p>The distinction from cognitive accessibility matters, although the two inevitably overlap.</p><p>Cognitive accessibility is concerned with barriers affecting areas such as attention, memory, language, reasoning, executive function, comprehension and task completion. W3C’s COGA work has built an extensive body of guidance around these needs. Its Working Group Note, <a href="https://www.w3.org/TR/coga-usable/"><em>Making Content Usable for People with Cognitive and Learning Disabilities</em></a>, is intended to make web content and applications more usable and accessible for people with cognitive and learning disabilities.</p><p>W3C’s work is also evolving. Its <a href="https://www.w3.org/TR/coga-research-modules/">2026 Cognitive Accessibility Research Modules</a> include work on online safety and wellbeing, decision-making, conversational interfaces and other circumstances affecting people who may require cognitive accessibility support.</p><p>Emotion is therefore not absent from cognitive accessibility. Stress can affect concentration. Anxiety can make decision-making harder. Emotional regulation can affect someone’s ability to use a service. Human cognition and emotion are intertwined, so an absolute boundary between them would make little sense.</p><p>The difference is primarily one of framing and emphasis. Cognitive accessibility examines barriers that affect a person’s ability to understand and use an interaction, including situations where stress or emotional state affect cognitive functioning. Emotional accessibility foregrounds the affective conditions created by the interaction itself, including its tone, pace, pressure and sense of safety, and asks whether those conditions constrain participation.</p><p>A form illustrates the distinction well. Clear language, predictable steps and manageable information can make it cognitively accessible. That same form may still create an emotional barrier if it uses threatening wording, artificial urgency or repeated warnings that make a routine interaction unnecessarily distressing.</p><p>This shifts some attention away from what the person can process and towards what the system is doing to create the circumstances in which processing takes place. The distinction is not an attempt to draw a hard psychological boundary between cognition and emotion. It is a design distinction: a way of bringing a different set of conditions into view.</p><p><strong>Inner experience and the first-person perspective</strong></p><p>There is another dimension to Neural Commons that becomes apparent when emotional accessibility is considered alongside research into inner experience.</p><p>Durham University’s Centre for Research into Inner Experience studies the subjective character of experience: how thought, perception, memory, emotion and other aspects of conscious life are experienced from the first-person point of view. Inner experience is broader than emotion, but that perspective has clear relevance to accessibility.</p><p>Much conventional digital evaluation privileges observable outcomes: whether someone completed a task, understood the information, encountered errors or abandoned the journey. Those measures are valuable, but they do not necessarily tell us what the interaction was like for the person experiencing it.</p><p>Someone may successfully complete a task while finding the experience intimidating. They may understand every instruction while finding the pace exhausting. They may reach the intended outcome despite feeling pressured or unable to participate on their own terms.</p><p>A first-person perspective can reveal aspects of experience that behavioural measures alone may not capture. This does not mean subjective accounts should replace established accessibility evidence. They can complement it, particularly when the subject being investigated is the emotional character of an interaction.</p><p>Bringing outward performance and inward experience into the same inquiry broadens the question of what meaningful participation involves.</p><p><strong>Exploring the emotional conditions of access</strong></p><p>Human-computer interaction, psychology, behavioural design, digital wellbeing and accessibility all examine parts of this territory. Neural Commons brings a particular set of those concerns together within an accessibility frame, with greater emphasis on first-person experience. It treats the emotional conditions of interaction as something that can be considered alongside more established dimensions of access, while asking how those conditions are experienced by the person on the receiving end.</p><p>That changes the kinds of questions designers can ask. Does a process give somebody sufficient time to think? Can they pause and return? Is urgency genuine? What happens when somebody chooses not to continue? Does the language explain a consequence accurately, or intensify it? Is slower interaction interpreted as failure simply because the product was designed around speed?</p><p>The white paper develops these ideas through proposed design commitments including emotional legibility, cognitive respect, trustworthy atmosphere, rhythmic flexibility, nonlinear entry and affect awareness. It also introduces the idea of “attunement metrics”, intended to explore whether factors such as a system’s tone, rhythm and atmosphere align with users’ needs alongside more established measures of accessibility and usability.</p><p>The work has subsequently developed into the <a href="https://www.bbc.co.uk/neurodiversity/emotional-accessibility-framework">Emotional Accessibility Framework</a>, translating the broader proposition into practical principles around psychological safety, nuance, context, reciprocity, adaptability and integrity.</p><p>The framework is beginning to move from a conceptual proposition towards questions that designers and researchers can investigate in practice.</p><p><strong>Pace as a condition of access</strong></p><p>Pace is particularly interesting because digital products constantly impose rhythm and that rhythm is rarely neutral. Sometimes it is visible, as with a countdown or an expiring offer. Elsewhere it appears through session timeouts, notifications, disappearing information, repeated prompts or expectations of an immediate response.</p><p>The Neural Commons paper discusses perceptual pacing, natural pause points and rest-positive interaction, alongside interfaces that allow people to slow down or move through information in less rigid ways.</p><p>None of this means that every interface should become slow or soothing. A fast interaction may be exactly what somebody wants. The accessibility question is whether speed, urgency and relentless engagement have been treated as unquestioned defaults, and whether people have meaningful ways to regulate their interaction when those defaults do not suit them.</p><p>Attunement becomes useful here because it shifts the question from how well a person adapts to the technology to whether the technology can accommodate variation in the person using it.</p><p>That may involve allowing more time, preserving progress when somebody pauses, providing alternative routes through information or avoiding pressure that serves no genuine functional purpose. The point is not to establish a universal emotional tempo for digital design. It is to recognise that tempo is part of design in the first place.</p><p><strong>When interaction becomes pressure</strong></p><p>The subject also intersects with a wider change in digital design. Many products are no longer passive tools waiting for input. They actively try to secure and maintain attention.</p><p>Notifications, engagement loops, recommendations, streaks and infinite feeds all influence when an interaction begins, how long it lasts and how easy it feels to stop. The design may be perfectly comprehensible while still exerting considerable emotional or attentional pressure.</p><p>Some of this territory is already addressed through work on deceptive or manipulative interface design. Article 25 of the <a href="https://eur-lex.europa.eu/eli/reg/2022/2065/oj?locale=en">EU Digital Services Act</a> prohibits online platforms from designing or operating interfaces in ways that deceive or manipulate people, or materially impair their ability to make free and informed decisions.</p><p>Article 25 also identifies practices such as repeatedly asking users to make choices they have already made and making termination more difficult than subscription as examples relevant to its prohibition on manipulative interface design.</p><p><a href="https://accessible-eu-centre.ec.europa.eu/content-corner/news/emotional-accessibility-and-how-it-relates-digital-service-act-2026-04-10_en?prefLang=el">AccessibleEU has connected Neural Commons and emotional accessibility with this regulatory discussion</a>, including practices such as constant notifications, infinite scrolling and artificial urgency.</p><p>Emotional accessibility has a somewhat wider scope. A design does not have to reach a legal threshold of manipulation before its effects become worth examining.</p><p>A system might be entirely honest about what it is doing and still be exhausting. It may interrupt too frequently, demand decisions too quickly or use a tone that makes a routine process feel adversarial. Those effects may be minor for one person and significant for another.</p><p>Treating them as part of access encourages us to look beyond whether an interface can be successfully operated and towards the quality of the conditions under which that operation occurs.</p><p><strong>Designing for variable human states</strong></p><p>This has obvious relevance to neurodivergence. Assumptions about the “right” pace, tone and sequence of an interaction are often embedded in digital products without being recognised as assumptions at all.</p><p>Frequent prompts may support one person and overwhelm another. A tightly linear process can feel reassuring to some users and restrictive to others. Hesitation may indicate confusion, but it may simply mean somebody needs more time before acting.</p><p>The Neural Commons white paper draws on feedback from a focus group of 45 neurodivergent and neurotypical participants alongside accessibility research and inclusive-design thinking. It also explicitly describes its assumptions about likely beneficiary groups as hypotheses rather than settled findings.</p><p>That exploratory status matters. The early work does not establish universal rules about how an interface should feel, and it would be premature to treat emotional accessibility as a completed standard.</p><p>At this stage, the value of the work lies partly in identifying questions that existing structures do not always make easy to ask together.</p><p>Its relevance also extends beyond diagnosis. People’s capacity to interact with technology changes with context. Fatigue, anxiety, unfamiliarity, grief, time pressure or a difficult day can alter what somebody needs from a service. A person who moves rapidly through an interaction on one occasion may need considerably more time on another.</p><p>Designing for variation recognises something ordinary about human beings: our emotional and cognitive resources are not constant, and neither is our inner experience of the systems we use.</p><p><strong>The measurement challenge</strong></p><p>Evaluation will be harder because emotional experience is contextual, culturally shaped and variable between individuals. A tone that reassures one person might irritate another. A pause that gives one user space may frustrate somebody trying to complete an urgent task.</p><p>Neural Commons proposes further work on attunement metrics and emotional-impact testing rather than presenting those methods as settled. Its policy proposals also consider how emotional tone and pacing might eventually form part of service-design evaluation.</p><p>The first-person perspective matters especially when it comes to evaluation. A meaningful evaluation of emotional accessibility is unlikely to rely on behavioural analytics alone. Completion rates, error rates and abandonment data can show what happened, but they cannot always explain the subjective conditions under which it happened.</p><p>First-person accounts, qualitative research and co-design therefore have a role alongside observable measures. The challenge is to combine them without reducing complex experience to a single score or assuming that one emotional response represents everyone.</p><p>Any move towards measurement, benchmarking or formal standards would require considerably more evidence, co-design and validation. The difficulty itself is not unusual in human-centred design. Concepts such as usability, trust and cognitive load also involve individual variation. They became useful because researchers developed better methods for investigating them rather than because everyone experiences them identically.</p><p>Emotional accessibility faces a similar research problem: identifying which aspects can be examined meaningfully without pretending that something as contextual as human emotional experience can be reduced to a universal measure.</p><p>Some aspects may eventually prove measurable. Others may work better as qualitative design principles, prompts for co-design or criteria for reviewing interaction patterns. An emerging field does not need to know all of its eventual methods before its central question becomes worth investigating.</p><p><strong>Why this framing is useful</strong></p><p>Problems without a clear conceptual home tend to scatter between disciplines. A stressful journey becomes a UX problem. Notification pressure is discussed as digital wellbeing. An overwhelming service becomes an accessibility issue. Manipulative wording moves to ethics or compliance.</p><p>Each field can see something important while the relationship between those problems remains harder to describe.</p><p>Emotional accessibility provides a way of looking across those boundaries. It allows researchers and practitioners to examine pace, pressure, tone, trust, agency, rest and different patterns of participation as related conditions of digital access rather than as entirely separate design concerns.</p><p>The inclusion of inner experience adds another useful dimension. It reminds us that an accessible interaction is not only something that can be observed from the outside. It is also something lived by the person moving through it.</p><p>The Neural Commons approach is particularly useful in bringing accessibility, affective conditions, neurodiversity, interaction design and first-person experience into the same conversation, and begins developing a language through which those relationships can be researched and designed for more systematically.</p><p>Neural Commons is developing an approach in which the emotional conditions of digital interaction can be researched, designed for and eventually evaluated with the same seriousness as other dimensions of access.</p><p>The framework remains open to research, refinement and challenge. It creates space for accessibility practice, neurodiversity research, human-computer interaction, research into inner experience and other fields to inform what emotional accessibility may eventually become.</p><p>Digital accessibility has often developed when barriers that were once diffuse became easier to identify, research and design for. Emotional accessibility extends that work to the affective conditions in which digital participation takes place.</p><p>The methods will need to mature, and the boundaries of the concept will continue to develop. The underlying design problem is already visible, however: technical and cognitive access do not necessarily tell us whether someone can participate meaningfully in a digital environment, or what that participation is like from the inside.</p><p>Neural Commons is helping make those conditions explicit enough to study and address in digital design.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f9c81f9dfa70" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The NHS App Is Becoming Part of Care. We Should Design It That Way]]></title>
            <link>https://medium.com/@leenahaque1/the-nhs-app-is-becoming-part-of-care-we-should-design-it-that-way-4c2568009734?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/4c2568009734</guid>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[ux-design]]></category>
            <category><![CDATA[health-technology]]></category>
            <category><![CDATA[healthcare]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Thu, 17 Sep 2026 11:19:09 GMT</pubDate>
            <atom:updated>2026-09-18T14:42:29.461Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Collage illustrating emotional accessibility in digital healthcare. A patient looks at a health app showing a new test result, referral under review, pending prescription, appointment and message, surrounded by handwritten questions expressing uncertainty. The calmer side reads “Digital care should feel like care” and shows supportive options including pausing, clear next steps, no action needed and returning later, highlighting pace, pressure, tone and trust." src="https://cdn-images-1.medium.com/max/1024/1*rPd7XIA402Ugu5SJuOYSyw.png" /></figure><p><em>What emotional accessibility could add to the next generation of NHS digital services.</em></p><p>Imagine opening the NHS App and finding a new test result.</p><p>You recognise none of the terminology. One number sits outside the normal range. There may be a clinician’s comment underneath it, or there may not. You don’t know whether the result is routine, whether somebody will contact you, or whether you are supposed to do something yourself.</p><p>The screen may be accessible, the text readable and the navigation working exactly as intended, while the person looking at it is still frightened.</p><p>This is becoming an important design question for the NHS because the NHS App is no longer a small administrative tool. By the end of July 2026, NHS England had recorded more than 42 million all-time account registrations, with more than 15 million distinct users logging in during July alone. <a href="https://digital.nhs.uk/data-and-information/publications/statistical/nhs-app-statistics/july-2026">NHS App Management Information — July 2026</a></p><p>The App is also becoming more deeply connected to hospital care. By April 2026 it was linked to every acute NHS trust in England, with around 64% of hospital appointments visible through it. Depending on the services connected to their account, patients can use it for test results, prescriptions, records, messages, referrals and appointment information. <a href="https://www.england.nhs.uk/2026/04/hospital-patients-can-now-check-appointments-in-the-nhs-app/">NHS England: hospital appointments in the NHS App</a></p><p>From 2027, that relationship is expected to deepen again. NHS Online will give people in England the option of receiving some planned specialist care through the NHS App, including online consultations, digital prescriptions and bookings for tests, treatments and procedures. Face-to-face care will remain available for people who need or prefer it. <a href="https://www.england.nhs.uk/digitaltechnology/nhs-online/">NHS England: NHS Online</a></p><p>As the App becomes part of the experience of being a patient, the quality of that interaction deserves the same attention as the functionality behind it.</p><h3>A test result is not an ordinary notification</h3><p>The NHS already recognises that test results need careful handling. Current NHS guidance allows clinical systems to control what patients can see in their records and to restrict information where there is a risk of harm. <a href="https://digital.nhs.uk/services/nhs-app/nhs-app-features/gp-health-records-in-the-app">NHS England Digital: GP health records in the App</a></p><p>Separate information-governance guidance addresses sensitive information more directly. Where information is likely to cause distress, such as a positive test result, staff are advised to ensure that an appropriate healthcare professional has spoken to the patient before it is added to the record. <a href="https://digital.nhs.uk/data-and-information/information-governance/guidance/access-to-patient-records-through-the-nhs-app">NHS England Digital: access to patient records through the NHS App</a></p><p>Those safeguards matter. They also illustrate why healthcare design cannot be reduced to whether information is available.</p><p>Consider the experience of waiting for a result. The patient may have been told that everything is probably fine. They may have been told very little. They may have spent the previous week imagining the worst. When a notification arrives, the meaning of that notification depends on everything that came before it.</p><p>A retail app can afford to treat “You have an update” as routine interface copy. In healthcare, the same phrase can land very differently.</p><p>Should a patient know before opening a result whether action is required? If a value sits outside a reference range, can the surrounding information explain what that means without encouraging the person to draw conclusions from a number alone? If somebody is confused or worried by what they are seeing, is the next step obvious?</p><p>These are design decisions. They influence what happens between receiving information and understanding it.</p><h3>Waiting is an interface too</h3><p>The same issue appears in referrals and hospital appointments.</p><p>Giving patients more visibility is useful. Patients at acute NHS trusts can now see referral and appointment information through the NHS App, with many also able to cancel or reschedule appointments. <a href="https://www.england.nhs.uk/2026/04/hospital-patients-can-now-check-appointments-in-the-nhs-app/">NHS England: hospital appointments in the NHS App</a></p><p>Visibility can also expose another part of healthcare that patients experience acutely: waiting.</p><p>A status such as “awaiting review” may be technically accurate while telling the patient almost nothing. Has somebody looked at the referral? Is the waiting time normal? Do I need to contact anyone? Will I lose my place if I do nothing? What happens if I hear nothing for another month?</p><p>A digital service cannot remove a waiting list. It can make waiting easier to understand.</p><p>There is a large difference between a status label and an explanation. Telling somebody what stage they have reached, what normally happens next and whether they need to take any action can reduce unnecessary uncertainty without offering certainty where medicine or capacity cannot provide it.</p><p>This is a practical example of what emotional accessibility can add.</p><h3>The conditions of access</h3><p>The Neural Commons Collective has been exploring emotional accessibility as an additional dimension of digital inclusion through its white paper, <em>The Right to Emotional Access in Digital Systems</em>. The work considers conditions such as pace, pressure, tone, trust and the experience of the person using a system. <a href="https://api.repository.cam.ac.uk/server/api/core/bitstreams/96130db0-b3e7-4d09-8205-773c444f5668/content">Read the white paper</a></p><p>Healthcare makes those conditions unusually visible.</p><p>A person booking a routine vaccination is in a different state from someone checking an oncology referral. The same design pattern may be unremarkable in one context and difficult in another. A countdown, warning, notification or repeated prompt carries different weight when the subject is somebody’s health.</p><p>An NHS service should not try to make every interaction reassuring. Sometimes information is serious. Sometimes action really is urgent. The important issue is whether the tone and urgency of the interaction are proportionate to the situation.</p><p>Urgent information should be unmistakably urgent when action really is required. Routine information should not become alarming because of careless language or visual treatment. Uncertainty should be explained where possible rather than hidden behind an unexplained status. People should be able to understand what is happening next.</p><p>That is a design problem as much as a content problem.</p><h3>Prescriptions show how small details matter</h3><p>Prescriptions offer a relatively simple example.</p><p>The NHS App lets people request repeat prescriptions and see information about their medicines. Prescription tracking is also being rolled out nationally, with participating pharmacies able to show whether medication is being prepared, is ready to collect or has been dispatched for delivery. <a href="https://digital.nhs.uk/services/electronic-prescription-service/national-prescription-tracking-service-for-patients">NHS England Digital: National Patient Prescription Tracking Service</a></p><p>By July 2026, more than 3,000 pharmacy sites were connected to the service, with almost half a million people viewing a tracking update during that month. <a href="https://digital.nhs.uk/data-and-information/publications/statistical/nhs-app-statistics/july-2026">NHS App Management Information — July 2026</a></p><p>This kind of visibility matters because an unexplained delay can generate anxiety quickly.</p><p>Imagine, for example, a status that said:</p><p>“Waiting for GP approval. You do not need to do anything at the moment.”</p><p>Compare that with:</p><p>“Pending.”</p><p>The administrative state could be identical. The experience of the person reading it is not.</p><p>NHS systems already use statuses such as Pending, Approved and Rejected for prescription requests, while prescription tracking can expose dispensing states such as preparing, ready to collect and dispatched. <a href="https://digital.nhs.uk/services/nhs-app/nhs-app-features/prescriptions-in-the-nhs-app">NHS England Digital: prescriptions in the NHS App</a></p><p>The opportunity is not simply to expose more system state. It is to make that state meaningful to the person who is waiting.</p><p>For somebody trying to work out whether they will have their medication tomorrow, a little context can be the difference between understanding the situation and spending an afternoon making phone calls.</p><h3>Accessibility standards are the foundation</h3><p>None of this reduces the importance of established accessibility work.</p><p>NHS digital services already operate within established requirements. NHS England’s current Accessible Information Standard guidance makes clear that websites and apps are governed separately by the Public Sector Bodies accessibility regulations and that organisations should meet WCAG 2.2 at level AA. The Accessible Information Standard itself addresses the information and communication support needs of disabled people and people with impairments or sensory loss. <a href="https://www.england.nhs.uk/long-read/accessible-information-standard-implementation-guidance/">NHS England: Accessible Information Standard implementation guidance</a></p><p>Those requirements address real barriers and remain essential.</p><p>Emotional accessibility asks another question once somebody can reach, perceive and understand the service: what is the interaction asking of them while they are there?</p><p>A form can have excellent keyboard accessibility and still create unnecessary pressure. Clear language can still sound accusatory. An appointment system can work technically while leaving somebody uncertain about whether anything is happening.</p><p>They are different kinds of problem, even when they occur on the same screen.</p><h3>Designing for the state people are actually in</h3><p>Healthcare has another complication. The same person is not the same user every time they open the App.</p><p>Someone who usually navigates technology easily may be exhausted after treatment. A person with no cognitive impairment may struggle to process information while frightened. Someone caring for a parent may be switching between their own needs and those of another person.</p><p>The NHS is also implementing the Reasonable Adjustment Digital Flag nationally. It is designed to make required reasonable adjustments visible across publicly funded health and social care, with providers required to be able to read, write and share this information by 30 September 2026. <a href="https://www.england.nhs.uk/learning-disabilities/improving-health/what-is-the-reasonable-adjustment-digital-flag/">NHS England: Reasonable Adjustment Digital Flag</a></p><p>That addresses enduring adjustment needs. There is an interesting question alongside it: how well can digital services accommodate needs that are situational or temporary?</p><p>A person may need to save a complicated process and return later. Someone receiving difficult information may benefit from a quieter route through it. A service could distinguish more clearly between a clinically urgent action and an administrative request so that the two are not given the same visual or linguistic weight.</p><p>The aim is to make fewer assumptions about the state of the person on the other side of the screen.</p><h3>The NHS is already changing quickly</h3><p>This is not a theoretical conversation about a distant health service.</p><p>The NHS App has recently been redesigned so that appointments, prescriptions and test results are easier to find from its home page. NHS England says the redesign followed feedback from patients and staff and included research with more than 1,000 users, including specific accessibility research. <a href="https://digital.nhs.uk/services/nhs-app/resources/new-look-nhs-app-what-is-changing">NHS England Digital: new-look NHS App</a></p><p>The wider direction is more significant still. NHS planning describes an App that increasingly becomes a single place for managing appointments, referrals and interactions, alongside a unified access model using AI-assisted triage to direct people towards the appropriate care. <a href="https://www.england.nhs.uk/long-read/medium-term-planning-framework-delivering-change-together-2026-27-to-2028-29/">NHS England: Medium Term Planning Framework 2026/27–2028/29</a></p><p>That shift is already underway. In July 2026, NHS England announced that a new AI triage tool in the NHS App was moving beyond trial. It is expected to reach more than 200,000 patients within 12 months and is planned to become available to all NHS App users by April 2028. <a href="https://www.england.nhs.uk/2026/07/nhs-accelerates-artificial-intelligence-rollout-to-cut-waiting-times-and-improve-care-for-millions/">NHS England: AI triage rollout</a></p><p>As those interactions become more sophisticated, their emotional character matters more.</p><p>An AI system does not simply display a menu. It asks questions, interprets responses and determines what appears next. Tone, pace, confidence and uncertainty become part of the interaction.</p><p>If an AI system tells a worried patient that their symptoms do not appear urgent, the wording matters. If it directs somebody towards emergency care, the wording matters. If it is uncertain, the way that uncertainty is communicated matters too.</p><p>Clinical safety, accessibility, content design and first-person experience begin to meet very quickly in a system like this.</p><h3>What an emotional accessibility review could look at</h3><p>There is no need to begin with a new national standard.</p><p>A useful starting point would be to take a handful of important NHS App journeys and examine them with patients in the circumstances in which those journeys actually occur: a test result, a referral, a prescription delay or a hospital message.</p><p>Researchers could look at whether people understood what was happening while also paying attention to where they felt uncertain, rushed, alarmed or stuck. They could identify moments where the service introduced pressure unnecessarily and see whether additional context, a different pace or a clearer explanation changed the experience.</p><p>The NHS already recognises the importance of involving users when improving digital accessibility. Its Accessible Information Standard guidance says organisations should involve end users, including people with different disabilities, impairments and sensory loss, when improving digital services. <a href="https://www.england.nhs.uk/long-read/accessible-information-standard-implementation-guidance/">NHS England: Accessible Information Standard implementation guidance</a></p><p>An emotional accessibility review could extend that enquiry into the conditions of the interaction itself.</p><p>That would be more useful than a general instruction to make healthcare technology feel “empathetic”. It would produce design evidence.</p><h3>Start with one journey</h3><p>The NHS has spent years bringing records, prescriptions, appointments, referrals and messages into fewer digital spaces. NHS Online will make the relationship between digital interaction and care deeper still.</p><p>That creates an opportunity to study these journeys differently.</p><p>Neural Commons offers one way of doing that. Emotional accessibility can help teams examine pace, pressure, tone, trust and first-person experience alongside established accessibility, usability and clinical-safety practice.</p><p>The useful work does not need to begin with a redesign of the NHS. It could begin with one journey: a test result, a referral, a prescription delay or a hospital message.</p><p>Work with the people who actually experience it. Find the points where uncertainty becomes distress, where a status stops being informative, where an interface creates urgency that the clinical situation does not require, or where a technically successful interaction leaves somebody unsure what to do next.</p><p>Then redesign that moment.</p><p>There is a worthwhile research and design question here for NHS teams, clinicians, accessibility specialists and people with lived experience: <strong>what would an emotionally accessible digital health service actually look like in practice?</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4c2568009734" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Neural Commons: Why Emotional Accessibility Matters in the Age of AI]]></title>
            <link>https://medium.com/@leenahaque1/neural-commons-why-emotional-accessibility-matters-in-the-age-of-ai-7d6b226b31ab?source=rss-4fa29c67eccc------2</link>
            <guid isPermaLink="false">https://medium.com/p/7d6b226b31ab</guid>
            <category><![CDATA[design]]></category>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[accessibility]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Leena Haque]]></dc:creator>
            <pubDate>Sun, 13 Sep 2026 15:38:57 GMT</pubDate>
            <atom:updated>2026-09-13T15:38:57.211Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="A woman sits thoughtfully at a desk beside a coffee mug and notebook, looking toward glowing, multicoloured streams labelled “Emotion”, “Attention”, “Values”, “Context”, “Intent”, and “Inner Experience”. The streams flow into a blue digital network on the right with image panels and the words “Personalises”, “Remembers”, “Recommends”, “Learns”, and “Adapts.” Large text at the top reads “Neural Commons”." src="https://cdn-images-1.medium.com/max/1024/1*A_ssWXOZesZz7kXqDWQpIA.png" /></figure><p>AI is changing the relationship between people and digital systems.</p><p>Interfaces increasingly personalise, remember, recommend, infer and intervene. They do more than help us complete tasks. They shape the experience around us.</p><p>That’s the territory I find most interesting in <a href="https://www.bbc.co.uk/neurodiversity/neural-commons/">Neural Commons</a>: the lived experience of adaptive digital systems, especially the way those systems affect emotion, agency, trust, attention and inner experience over time.</p><p>The early thinking is set out in the white paper <a href="https://api.repository.cam.ac.uk/server/api/core/bitstreams/96130db0-b3e7-4d09-8205-773c444f5668/content">The Right to Emotional Access in Digital Systems</a>, developed through work involving BBC CAPE, Cambridge ThinkLab and Durham University’s Centre for Research into Inner Experience.</p><p>Sean Gilroy has also written about <a href="https://www.linkedin.com/pulse/bbc-neural-commons-what-emotional-accessibility-digital-sean-gilroy-hfeje">BBC Neural Commons and emotional accessibility in digital design</a>. His piece looks at the design implications of emotional accessibility. I want to focus on a related question: what happens when the system itself becomes an active participant in the experience?</p><p>For most of the web’s history, software was relatively passive. You opened a site or application because you wanted to do something. You searched, clicked, entered information, made a choice and left.</p><p>Adaptive systems behave differently.</p><p>A recommendation system influences what appears next. A notification decides when to interrupt you. A conversational AI can affect tone, pace and perceived authority. Personalisation can alter which options you encounter. Memory changes the relationship over time. Inference allows a system to act on what it thinks it knows about you, whether or not you’ve explicitly said it.</p><p>That changes the design problem.</p><p>The experience is no longer defined only by the interface in front of you. It’s also shaped by what the system remembers, predicts, withholds, recommends and assumes.</p><p>This is where I think Neural Commons becomes distinctive. Its unit of analysis isn’t simply the interface, the user or an individual interaction, but the evolving relationship between a person and an adaptive system over time.</p><p>Attention, emotion, trust, agency and cognition are often discussed separately. Neural Commons asks what happens when they interact inside systems that are increasingly capable of changing their behaviour in response to us.</p><p>That relationship becomes a different object of study. It’s not only about whether someone can use a system at a particular moment, but what happens as the system learns, adapts and intervenes across repeated interactions.</p><p>A small example is personalisation.</p><p>Personalisation is often presented as an obvious good. If a system can adapt its complexity, pacing or presentation to somebody’s preferences, that could remove genuine barriers.</p><p>But adaptation can also become assumption.</p><p>There is a meaningful difference between changing a system because you want it to work differently and discovering that the system has quietly decided what it thinks is best for you.</p><p>The first can increase agency. The second can make agency harder to see. That distinction becomes more important as systems get better at inference.</p><p>If AI becomes increasingly capable of deciding what we are likely to want, the person needs to remain capable of correcting the system, changing their mind, seeing how decisions are being made and reversing them.</p><p>A system can be personalised without being genuinely under the person’s control.</p><p>The same is true of trust.</p><p>Generative AI can be fluent and wrong at the same time. It can sound confident when the underlying answer is uncertain. Conversational systems use language that people associate with empathy, expertise and authority, even though the system does not possess those qualities in the ordinary human sense.</p><p>The user therefore has an extra task while interacting with AI: judging the system itself.</p><p>Is this answer reliable? Should I verify it? Why has this recommendation appeared? What has influenced the response? What does the system remember about me?</p><p>That interpretive work takes effort.</p><p>Some people may barely notice it. Others may find that uncertainty much harder to manage, particularly when consistency, predictability or literal interpretation matter to them.</p><p>A system can be easy to operate and still be difficult to trust. That’s where emotional accessibility becomes useful.</p><p>For me, emotional accessibility asks whether a digital system creates avoidable emotional demands that make it harder for someone to participate, understand what is happening, exercise agency or remain comfortable in the interaction.</p><p>It gives us a way of examining aspects of an experience that conventional product metrics can easily miss.</p><p>Someone can complete a digital journey successfully while feeling hurried throughout it. They can understand the content while finding the interaction exhausting. They can remain engaged with a service because the service offers few natural stopping points.</p><p>The dashboard may record success. The person may experience something very different.</p><p>Emotional accessibility asks us to notice that gap.</p><p>Was there unnecessary urgency? Could the person pause without losing their place? Did the system make uncertainty clear? Did personalisation help, or did it make the experience feel less controllable? How much effort did the person have to spend staying comfortable, oriented and confident about what was happening?</p><p>These questions matter even more when adaptive behaviour accumulates over time.</p><p>We may eventually stop thinking consciously about “using AI” in the same way we no longer think about “using the internet”. It will simply be present in search, work, education, healthcare, entertainment and public services.</p><p>At that point, the cumulative experience matters.</p><p>A single notification may be trivial. A recommendation may be useful. A personalised intervention may save time. But repeated across days, months and systems, these interactions can begin to shape attention, expectations and decision-making.</p><p>Emotional accessibility therefore also changes how we think about attention. If a system can increasingly predict what will keep someone engaged, accessibility cannot be judged only by whether someone can continue interacting. We also need to ask how easy the system makes it to stop.</p><p>Much of the digital economy has been built around capturing attention. Infinite feeds, autoplay, alerts and recommendation systems all reflect the assumption that more engagement is usually better.</p><p>AI could make that logic far more sophisticated. A system that understands patterns in behaviour can become extremely good at predicting what will hold somebody’s attention.</p><p>The question is whether that should always be treated as success. I’m more interested in whether the interaction continues to deserve the person’s attention. That may sound like a small distinction, but it changes the design objective.</p><p>Instead of asking how to keep somebody engaged for longer, we can ask whether the interaction remains useful, understandable and under their control.</p><p>That shift begins to move us from an attention economy towards something closer to a trust economy.</p><p>Trust is harder to measure than time spent, but it may tell us much more about whether the technology is actually serving the person using it.</p><p>There is another design assumption worth questioning too: friction.</p><p>Digital design has spent years trying to remove it. Fewer steps, faster decisions, automatic continuation. Much of that has improved experiences. But some friction protects agency.</p><p>A pause can give someone time to reconsider. A confirmation can prevent an unintended action. An explanation can make a decision easier to understand. Being asked can sometimes be more respectful than being predicted.</p><p>The interesting question is not whether friction is good or bad. It is whether the friction obstructs somebody or gives them more control.</p><p>Adaptive AI makes that distinction more important because these systems increasingly act before we explicitly ask them to. This is the research space Neural Commons is beginning to explore. It’s not simply about how people respond to technology, but the lived relationship between people and systems that increasingly respond to them.</p><p>That relationship includes emotion, attention, trust, agency, personalisation, uncertainty and inner experience because people experience those things together.</p><p>There are still difficult questions to answer.</p><p>Where do adaptive systems create avoidable emotional demands? When does personalisation become intrusive? How should systems communicate uncertainty? How do different people experience memory, interruption and recommendation over time? What does meaningful agency look like when a system is continuously adapting?</p><p>And how do we study these experiences without deciding in advance what somebody ought to feel?</p><p>Those are empirical questions. The answers need evidence, lived experience and careful research.</p><p>As technology becomes better at adapting to people, we need to make sure people don’t end up spending even more effort adapting themselves to technology.</p><p>That, to me, is what emotional accessibility is asking us to consider. And it’s why Neural Commons matters now.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7d6b226b31ab" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>