Issue #08 is only 5 pages long, front cover and back pages included. This week we took a short break to rest, so this issue is short with reminders of our purpose and mission as a company.

As always, here’s a link to the PDF version to download and read.

Let’s get to it!

Page 1: Qualitative research: An interaction with my 4-year-old cousin reminded me to check my bias

A few Sundays ago my baby cousins (4 and 2) came to visit at my parents house. They do this almost every Sunday together with their mum and dad. But this Sunday, they only came with their dad. After all the greetings and pleasantries, I asked my 4-year-old cousin where his mummy was. In his words “mummy is in the shop”. His mum owns a salon. Then I asked, “what is mummy doing in the shop?”. He replied “mummy is drawing” while pointing at his wrists. You probably have an idea what he meant by this, but because I knew his mum’s salon to provide only hair and mani-pedi services, I assumed he meant she was doing someone’s nails and by drawing he was referring to some acrylic design. I didn’t wonder why he was pointing at his wrists, I just thought nails, wrists, it’s all hands and let it go. It wasn’t serious.

By evening time when his mum was done at work and was now at our house, she asked to show me what she did at work today. Lo and behold, it was a tattoo — a wrist tattoo! Apparently she had added a tattoo parlor to her salon and I obviously didn’t know this. Instantly, I thought of my earlier interaction with my cousin and how I had completely misunderstood him, even tried to correct him, meanwhile he was the one who was right. Mummy was indeed drawing!

From this interaction I was reminded of how biases, in this case, an innocent confirmation bias could taint the outcome of a research affecting things like policies, insights, etc. The thing about biases is that they are most times very unrecognizable when being exhibited. I assumed what my cousin meant and concluded on that, based on the knowledge I had, instead of leaving room to understand certain things like why he kept pointing at his wrists and why he repeatedly said no when I replied ‘oh, she’s painting nails’. I assumed because he is younger and basically a toddler, he didn’t know what he meant but I did. I subconsciously deemed him unreliable.

This was just one tiny interaction but an example of how confirmation bias works, is present in all of us, and could affect the end result of a research project if not checked. So this is a reminder to you as a qualitative researcher to trust your research participants and instead of concluding or assuming, maybe ask, maybe spend more time trying to understand their point of view.

Page 2: A company and a design project similar to our goals at DIAL Design

1. Greater Good Studio: A design firm dedicated to the social sector

Founded by Sara Cantor and George Aye, Greater Good Studio is a human-centered design firm dedicated to the social sector. The. company refer to themselves as a ‘social enterprise, both mission-driven and for-profit’. GGS works alongside people impacted by social inequities to co-design strategies and solutions that are inclusive, innovative, and impactful.

2. Accessible voting system designed by IDEO:

In 2015, Los Angeles County, the largest jurisdiction in the U.S.hired IDEO, a design company to create a more accessible voting experience to accommodate everyone including non-English speakers, the hearing- or vision-impaired, people who use wheelchairs or other mobility devices, and those who aren’t tech savvy.

After extensive research with diverse constituents of the country, IDEO was able to design a solution that makes voting easier for everyone. This solution included. a self-enclosed ballot-marking device with a touchscreen that allows any voter to easily navigate, mark, verify, and submit a paper ballot. It also included audio and tactile options for those with vision or reading impairments as well as the option to read ballot in any of the over 10 languages available. IDEO’s solution didn’t stop at the device but also extended to the whole voting process including vote-by-mail ballots, interactive sample ballots, and the tally process.

Page 3: About DIAL Design

Design is an integral part of our daily lives, shaping everything from technology and transportation to cities, communities, and more.

However, designs can perpetuate discrimination, exclusion, and harm against certain groups. Historical examples include car seat belts designed without considering women’s bodies, resulting in disproportionate harm, and Face ID systems that struggle to recognize darker skin tones or those with facial differences.

The Problem

Biased designs cause frustration, endangerment, or even life-threatening consequences for marginalized groups. To address this issue, we must revisit the design process’s foundation which includes data, research, and practices.

Our Mission

At DIAL Design, we recognize the impact of biased design on individuals and businesses. We work at the intersection of design equity and social justice and exist to help mission driven companies and the public sector create ethical, inclusive experiences across various industries. Our goal is to promote equity-centered design through:

• Better Data: Ensuring data collection, processing, use, and curation are representative and unbiased.
• Inclusive Research: Conducting research that considers diverse perspectives and needs. Especially those of the people who are most likely to use the products or be impacted by it.
• Accessibility: Designing products and services that are usable by everyone.
• Localization: Adapting designs to accommodate cultural and regional differences of underrepresented communities.

Our Values

• Equity: prioritizing fairness and justice in design practices.
• Respect: valuing people as more than statistics and users but as humans with diverse perspectives and experiences.
• Co-design: working together, especially with solution providers, designers, and members of communities that will be impacted by designs, to build inclusive solutions.
• Transparency: we practice open and honest communication with partners, clients, communities, etc

Contact Information

LinkedIn: DIAL Design

Email: dialdesign26@gmail.com

Founder’s contact: ebosejenna@yahoo.com/ Ebosetale Jenna Oriarewo on LinkedIn.

Page 4:

On this page is a graphic of a person on a clock and it reads ‘we are on a short break but we’ll be back with more insights next Monday’.

Page 5: Thank you for reading

DIAL Weekly is a weekly magazine run by DIAL Design. Every Monday 10am EDT, we share an issue containing design equity insights to help us think and approach design more equitably.

To find out more about us or our services, visit our LinkedIn (DIAL Design) or email dialdesign26@gmail.com.

Signed: Ebosetale Jenna

DIAL Weekly #08 was originally published in DIAL Design on Medium, where people are continuing the conversation by highlighting and responding to this story.

Another week, another DIAL Weekly. If you prefer a pdf version, you can download/read issue #07 using this link.

This issue was centered on accessibility outside of digital technologies. In the total 9 pages, we look at companies providing services enhancing accessibility in the physical world, spotlight an accessibility educator and the Marsham Court Hotel for its accessibility efforts, and more.

Let’s get right into it!

Page 1: Accessibility in physical environments.

Accessibility is more than a digital issue, it also is a physical and environmental one. If disabled people exist in that sphere, then accessibility should be an important part of the development process and experience.

In April of this year, we set out to understand the ways blind and visually impaired people are impacted by inaccessibility in both digital and physical spaces. So we ran a survey (as well as a focus group session). While our survey covered the African experiences of this particular group, we only included responses from those in Nigeria when we worked on the minimal version of the report.

Question 8 of the survey asked “What’s something you’ll love to see fixed to be more inclusive of visually impaired people? The most common responses we gathered were about the physical environment such as roads/walkways, ATMs, POS machines, elevators, and someone even used the words ‘’physical environments and public spaces”.

Some other non-digital responses we got included:

• Sporting activities
• Public transportation. Implementing audio cues, braille signage and tactile indicators on platforms and vehicles
• Braille sign on currency notes

etc

Read the full report here.

It is important to note that aside infrastructural changes like these mentioned here, accessibility isn’t just the responsibility of the government, you as an individual could also impact accessibility in your space. Whether it’s by building a workplace that is accommodative of those with disabilities or even a party. Whatever it is, accessibility is a collective responsibility and in this issue, you’ll learn about implementing it in the physical environment.

Page 2: WelcoMe and Sociability

WelcoMe helps you improve the physical shopping experience of your disabled customers

WelcoMe is a Scottish based web app developed by former guide dog trainer Gavin Neate to help retailers provide tailored and more accessible services to disabled customers.

Currently implemented across 160 venues in the UK, the app works by letting disabled customers schedule a visit including their time of arrival and the specific support they require. With this knowledge, retail staff can then adequately prepare ahead of time on how to serve them.

In the UK, approximately 16 million (1 in 4) people are disabled with a total spending power of around £274 billion annually. Improving the retail experience of people with disabilities therefore, matters on both a human rights level and an economic one. With improved customer service, businesses increase sales, brand loyalty, and customer satisfaction.

WelcoMe’s goal is to provide a more equitable and inclusive physical shopping experience for disabled people by empowering staff with knowledge and also customers with the power to reduce the barriers they experience.

Discover accessible venues with the sociability app

For disabled people, sometimes finding public spaces that are built to support your needs can be difficult. Whether it’s a cafe with an accessible bathroom or a church with provisions for deaf and hard of hearing worshippers or parks with disability-friendly playground equipmen for kids. Sociability solves this problem by providing easy access to detailed information about accessible spaces. They do more than tag a place as accessible or not, they provide specific details on what each space features so you can select the one that’s just right for your needs.

The sociability team is made up of people with a varying range of disabilities who bring their lives experiences and expertise to their mapping exercises. They also work with members of communities to find and map locations. Currently they have mapped over 14,000 venues in the UK and have almost 15,000 users.

“Too often disabled people are left spending hours ringing around and not finding the answers they need or wandering around trying to find somewhere to go that can accommodate their needs. Couple that with last-minute plans and it’s usually a recipe for disaster.

Using Sociability you can search for bars, café’s, restaurants, and pubs, and filter using the specific access requirements you may have. We capture a whole host of information, from mobility-related accessibility to visual, hearing and sensory.”

Page 3: Accessible Events 101: 6 tips for organizing accessible physical corporate events with pointers from the 2024 Democratic National Convention

1. Provide live audio descriptions: Audio descriptions are beneficial for blind and low vision people to help them understand what exactly is happening in a place or on tv. While this was usually available after the conference, this year it was available live and also in Spanish.
2. Ensure live captioning or/and a sign language interpreter are available: The convention was streamed live with American Sign Language interpretation available as well as live captioning to aid deaf or hard of hearing attendees and viewers.
3. Sitting & floor plan arrangements should be inclusive of wheelchair users: Unlike other years where delegates in wheelchairs had a sitting area separate from the other delegates, this year a ramp was added to grant access to the convention floor so all delegates could sit together regardless of abilities.
4. Sensory accommodations should be included for those with sensory issues: The DNC provided sensory kits included things like noise canceling headphones to help delegates and attendees who had sensory issues like autism. Sensory or quiet rooms with low lighting are also a provision.
5. Accessible transportation (which also includes disability parking): Wheelchair-accessible buses were provided and available all through the convention. In addition, scooters were available to borrow at the convention for those who couldn’t be on their feet for long periods.
6. Trained support should always be available: For those who needed it, there was support and assistance available to them once they arrived at the airport from someone who had been briefed on their specific needs. Ask guests prior if they’ll require disability support/assistance of any kind and make that available to them.

Page 4: Meet the woman helping make bathrooms and spaces more accessible — Julie Sawchuk

Julie Sawchuk is a Canadian-based Accessibility Educator and Strategist helping make spaces more accessible through her company Sawchuk Accessibility Solutions. She’s especially popular for her #ToiletTuesdays LinkedIn posts where she as a wheelchair user audits public bathroom and toilet spaces she encounters to determine whether or not they are accessible for wheelchair users or people with mobile disabilities.

She is also a best selling author of Building Better Bathrooms. Her journey in accessibility began after an accident in 2015 which was when she became a wheelchair user. After a 2-year redesign of her home to be wheelchair accessible, she realized how much more efficient the changes made her. Since then she has built her company educating and consulting with designers on home to public space renovations, become a. Rick Hansen Foundation Accessibility Certification Professional, and much more. You can find her at www.juliesawchuk.ca

The image includes the words in this section as well as a a photo of Julie — a white woman with blonde hair, she’s wearing glasses and is seated legs crossed in her wheelchair. Beside her photo is another photo, this one is of her book.

Page 5: Accessibility at the Marsham Court Hotel

Marsham Court Hotel is a family owned hotel located in Bournemouth, England. The building is over a 100 years old and yet, one of the most accessible hotels in the UK. It’s a 95 room full service hotel whose current management are especially keen on accessibility and inclusion. This is evident not only in their efforts towards hiring diverse staff members, but also in their renovated additions/provisions of accessible facilities. Including 5 rooms with profiling beds and removable ceiling hoists, some of which are dog friendly and also have connecting double-bedded rooms ideal for disabled guests with carers or family members.

Located in a town that sees about 5 million tourists annually that also has 73,838 disabled people according to a 2021 census, investing in accessibility provides the hotel with an advantage over many others. By catering to the needs of disabled people and their carers — a market globally estimated at about $18 trillion USD. and £274 billion in the UK — the hotel increases their customer pool while also building brand loyalty with people.

The hotel also has a sensory room for visitors and residents with sensory issues. Installed in November 2021, the sensory room was designed to help people relax and de-escalate. It is also open to residents and non-residents, although non-residents may be charged a £5 fee.

Only hotel in the UK with a changing place facility

A changing place refers to a fully accessible toilet and bathroom space fitted with amenities that cater to the needs of those with a varying range of disabilities, especially those that can be impacted by the quality of the bathroom or toilet space. In an episode of DHM’s — Hotel Digital Marketing podcast, Rosie Radwell, Managing Director of the Marsham Court Hotel mentioned how they were able to fund this project through a grant funding of £70,000 and till date they are the only hotel in the UK with this facility in place.

Some of the facilities provided in the Marsham Court Hotel’s changing place include:

• Colostomy shelf
• Height adjustable wash basin
• Nonslip flooring
• Wall mounted height adjustable, adult-sized changing bench
• Ceiling hoist
• Peninsular toilet with drop-down grab rails on each side
• Privacy screen
• Shower chairs of different types

When thinking about investing in accessibility, a worry for most companies is the return on investments. Following this hotel’s efforts, not only has this opened them up to a new community of repeat customers, their marketing is heavily from word of mouth from satisfied visitors as well as influencers.

Page 6: 5 concluding points on how to approach accessibility.

Useful whether you’re designing physical products, digital products, or services.

1. When in doubt, start with data:

Collect data from your users on their disability status or if there are any existing or new functionalities that will improve their user experience.

Also, consider studying existing data and research on the disability market value generally and for your industry.

2. Begin with smaller changes:

Understand how assistive tech works and the accessibility provisions you can begin making. from now eg adding alt texts to images, captions/subtitles to videos and audio content, provide training for your staff or design team.

You can also adopt an accessibility guideline like the WCAG and follow its provisions.

3. Run product tests with diverse categories of people:

If accessibility is a goal for you, then you need to approach it with the understanding that reading about it isn’t enough, you also need to include disabled people in the development and testing process of your product or service.

4. Support accessibility across your product/service. Not only one part:

Accessibility should be a comprehensive effort impacting every customer touchpoint. Improving your website while your stores or offices or social media or marketing or customer support efforts remain inaccessible leaves a gap. A good user experience design caters to every stage of the customer’s journey. So should it be with accessibility.

5. Consider the simple and true fact that accessibility grants people access and is the right thing:

Change your mindset on accessibility as a special feature and instead see it as a right. Then it would be easier to make it a part of your product. Accessibility features benefit everyone from ramps. to subtitles to text magnifiers. It improves overall user engagement, retention, and experience and it opens you to more customers.

Page 7: Webinar: Disability Literacy 101 — Inclusion Education for the Workplace

This webinar held yesterday the 20th of November and was organized by the Inclusive Workplace and Supply Council of Canada (IWSCC) and presented by Mindset Executive Producer Jeff Tiessen, PLY. Informed by 45+ years of lived-experience as a double-arm amputee, and 30+ in the media business, Tiessen is a 3x Paralympian and gold medalist, award-winning author, advocate, publisher, producer and educator.”

Page 8: Get started on accessibility with DIAL Design

From accessibility-focused research, recruiting testers, accessibility audits, accessible content creation, and compliance, we at DIAL Design help ensure your products and services are usable for all people regardless of their abilities or disabilities.

Contact us: dialdesign26@gmail.com

Page 9: Thank you for reading!

DIAL Weekly is a weekly magazine run by DIAL Design. Every Monday 10am EDT, we share an issue containing design equity insights to help us think and approach design more equitably.

To find out more about us or our services, visit our LinkedIn (DIAL Design) or email dialdesign26@gmail.com

Signed: Ebosetale Jenna.

DIAL Weekly #07 was originally published in DIAL Design on Medium, where people are continuing the conversation by highlighting and responding to this story.

As usual, if you’ll prefer a pdf version, here’s a link to issue 06. This issue is 8 pages long, including the front and back page. We focus on lexical racism, raciolinguistics, and language bias both in the real world and now in AI tools. Here’s all of it.

Front page: Lexical racism and language bias from real world to AI

If you agree systemic inequities exist in society, especially towards certain groups of people in the form of racism, sexism, ableism, etc. Then it is important to acknowledge that these inequities make their way into human designed products, services, and systems. Intentional or not.

Biases are present in every society and they exist in various forms both implicitly or unconsciously (things we don’t realize we are judging a person by) as well as explicitly or the more outwards forms of bias. As many discussions around the development of AI and AI ethics have mentioned, these biases are highly likely to show up (and even currently present) in existing and even future models further perpetuating exclusion, discrimination, and sometimes harm. With this, many have placed a big focus on data and algorithm audits to mitigate biases. While this is applaudable effort, recent reports show that a lot of these mitigation efforts only solve the explicit biases found while implicit biases remain and are even amplified.

“Previous AI research has revealed bias against racialized groups but focused on overt instances of racism, naming racialized groups and mapping them to their respective stereotypes, for example by asking language models to generate a description of a member of a certain group and analysing the stereotypes it contains. But social scientists have argued that, unlike the racism associated with the Jim Crow era, which included… (continued in page 4 & 5).

Page 2:

There are 2 subheadings on this page.

First heading: In 2013, the court rejected Rachael Jeantel’s testimony during the George Zimmerman trial because her speech delivered in Haitian-influenced AAE was considered uneducated and untrustworthy.

19-year-old Jeantel was the last person to speak with 17-year-old Trayvon Martins few minutes before he was shot and murdered by police officer George Zimmerman in February of 2012. Jeantel was therefore a star witness in this case and spent about 6 hours delivering testimonies about that night. Her testimony would end up getting thrown out after online comments, news outlets, and even the jury deemed her unreliable and untrustworthy, with some calling Jeantel an illiterate. All simply because she spoke in a dialect they didn’t speak, couldn’t understand, and had prior attached stereotypical biases to.

Jeantel who is trilingual (she speaks Spanish, Haitian Creole, and English but in a HaitIan African American Vernacular) had her deposition and witness. testimony stigmatized because not only could the jury not understand her, the court transcriptions contained a lot of error, according to Stanford linguists John Rickford and Sharese King. Additionally, tests performed by linguists on court reporters in Philadelphia — the city of brotherly love — showed that despite being certified at a 95% accuracy level, when faced with Black English, they made errors 40.5% of the time. It also doesn’t help that according to a University of Pennsylvania published paper, juries are exceptionally biased towards speakers of Black English, even when they can’t see them.

Second heading: Warren Demesme was denied due legal process because the courts thought he was requesting a lawyer dog

In what is known as the infamous ‘lawyer dog’ case of Louisiana, in 2015, Warren N. Demesme was facing accusation of sexual assault from 2 underage girls. While being questioned by the police, a transcript of the interview showed that Warren told the police: “This is how I feel, if y’all think I did it, I know that I didn’t do it so why don’t you just give me a lawyer dog cause this is not what’s up.”

The Louisiana court in response voted 6–1 to hearing his appeal and providing him legal counsel. According to one of the state’s judges Scott J. Crichton, Warren’s request for a ‘lawyer dog’ was ambiguous. In his words. “In my view, the defendant’s ambiguous and equivocal reference to a ‘lawyer dog’ does not constitute an invocation of counsel that warrants termination of the interview.”

Shocking that not one person thought to confirm the validity of a sane adult requesting a lawyer dog. Because the transcription missed a comma between lawyer and dog (dawg which is an AAE slang for friend/brother), Warren remained without access to his Sixth Amendment Right even though he had placed a request, albeit in a dialect not one member of the legal process claimed to understand.

In Louisiana where one-third of the residents are Black, one would expect a certain level of awareness or knowledge of AAE, especially in the state’s highest courts. But sadly, that isn’t the case.

Page 3: 2 Culturally Inclusive Language Models Built For & With The Black Context

Language Models are trained on very large datasets — content — scraped from all over the internet. Content which carries with it biases and are sometimes incomplete and non-diverse. Millions of people use these language models in some form today across industries. From creating learning plans, to writing, to translations to informing hiring decisions and more. When certain contexts are missing from a tool universally as important as language models are, certain groups are left with majorly negative experiences that exclude them from reaping the benefits of such innovation.

Here are 2 models built to cater to black contexts:

1. ChatBlackGPT is a chatbot just like ChatGPT only with more insights into and from the Black perspective. ChatBlackGPT was founded by Erin Reddick an AI Culture Technologist and was developed as a community effort with the team including students from historically Black colleges and universities (HBCUs) who helped teach and train the tool while also providing a learning opportunity for Black and Brown people to safely explore the AI space. Reddick in a TechCrunch interview stated the algorithm prioritizes Black information sources so it’s immediately relevant and relatable to the average Black experience.

2. Canada-based and founded by Tamar Huggins, Spark Plug is a platform that translates educational materials into African American Vernacular English (AAVE). Huggins trained her algorithm on the verbiage of her teenage daughter as well as Black authors in education and from the Harlem Renaissance writings. She consistently partners with linguists, educators, and cultural experts to audit the inputs and outputs produced by Spark and unlike with ChatGPT, users are in control of their data. Huggins is ensuring Black kids can see themselves in education and find simplified processes to help them stay engaged and achieve success.

Page 4: continued from page 1

…overt behaviours such as name calling or more brutal acts of violence such as lynching, a ‘new racism’ happens in the present-day United States in more subtle ways that rely on a ‘colour-blind’ racist ideology. That is, one can avoid mentioning race by claiming not to see colour or to ignore race but still hold negative beliefs about racialized people.”

This quote and the attached image above are from a research paper titled “AI generates covertly racist decisions about people based on their dialect” by Valentin Hofmann, Pratyusha Ria Kalluri, and Sharese King.

Implicit biases are those biases we likely don’t recognize we hold about certain groups. They aren’t even always harmful, but they feed into stereotypical ways of approaching people. A good test of how much implicit bias you have is reading ‘Recitatif’ by Toni Morrison — a short story with no mention of race but we end up assigning races to the characters based on our inherent biases.

While implicit biases are a thing that affects different groups of people, this article specifically focuses on Black people and the assigning of intelligence or intellectual worthiness simply based on the use of African American English (AAE) or African American Vernacular English (AAVE).

The following are from a LinkedIn post I made last week.

“There’s a lot of racism in the ways people attribute intelligence.

Recently, Cardi B, an American rapper endorsed Kamala Harris for president while speaking at a Kamala Harris rally in Milwaukee, Wisconsin. While delivering her speech she spoke in her regular dialect which Cardi herself calls a broken form of English and could be seen as a form of African American vernacular. But because of how she speaks, there were comments that called her unintelligent, incoherent, and ‘clearly not educated to speak simple English’.

I know there are other (very silly) factors that make people think she isn’t a worthy person to endorse a political candidate, but some of what I’ve seen, maybe half of the comments were about her dialect and the way she speaks.

This is an example of the accent/dialect prejudice that exists with people attributing certain biased traits to people based on how they speak. And like every other unchecked racial stereotypes making its way into AI, so is this one.

From machines not understanding the voices/accents of non-Westerners or those with speech differences to AI now furthering these harmful racial biases based on people’s dialects alone, again, we need more intersectional algorithm checks. I say intersectional because while there’s a lot of focus on the more blatant forms of racism, there’s not much about the hidden forms.

We also need to un-popularize the scraping of and use of unchecked data and instead emphasize the importance of model developers collecting and building more diverse datasets.”

Page 5:

“From dialect to linguistic abilities, it is clear that language bias is a thing making its way rapidly across technology that is used in legal systems, education, housing, hiring, and just about every other sector. And the implications for those who don’t fit the colonial standard — white — are clearly dire.”

The thing with language biases and lexical racism is that it extends more than spoken dialect, it also permeates written language. Earlier in the year, venture capitalist Paul Graham shared on Twitter how words. like ‘delve’ are an indicator of AI use because ‘it’s one of those words like ”burgeoning” that people only use when they’re writing and want to sound clever.’ This sparked an entire conversation, especially on Nigerian Twitter with Nigerians explaining how normal of a word delve is and even providing examples of how Nigerian English specifically is ‘‘flowery, musical, emotive, dramatic”. From newspaper headlines to videos and many more, Nigerians declared a loud and proud war against what gradually became a heated conversation with some blatant racism getting involved. Like people asking how Nigerians who they think aren’t educated can know such words.

While Graham hadn’t singled out any race or country in his tweet, it became an important conversation to be had and defended by Nigerians who have been recipients of racism from hiring algorithms, faced accusations from lecturers abroad about cheating on assignments, white colleagues offering to rewrite emails they thought were too eloquent for an Nigerian and thus must be AI-written, and as data even shows AI detectors are biased against non-native English writers. From dialect to linguistic abilities, it is clear that language bias is a thing making its way rapidly across technology that is used in legal systems, education, housing, hiring, and just about every other sector. And the implications for those who don’t fit the colonial standard — white — are clearly dire.

A comment under the LinkedIn post I shared in the previous page asked: ‘It’s disheartening to see how deeply ingrained accent/dialect prejudice is, even manifesting in AI biases. The Cardi B example highlights this perfectly her eloquence and intelligence are overshadowed by the way she speaks. I mean, it’s like the system prioritizes a narrow standard of “proper” English, disregarding the richness and nuance of other dialects. Given the increasing use of AI in education and hiring, how can we ensure these algorithms don’t perpetuate existing inequalities based on linguistic diversity?’

The short version of my response was asking for regulations on algorithms. Whether this is regulating the areas/use cases/industries where they are used like the NY Law 144 regulating the use of AI in hiring processes and the EU AI Act that regulates the development of algorithms to be used in certain sensitive areas. Or it’s by in-house governing the development of these algorithms across each stage from data collection to labeling to model training to deployment. Bias gets into these models through any of these stages but can be minimized or eliminated by being intentional in your approach towards mitigation.

At DIAL Design, we recognize the complexities of developing bias-free algorithms. That’s why we provide strategies and services that help streamline the process.

From inclusive data collection to a more collaborative intersectional approach to model training and algorithm audits by involving a diverse blend of potential users in the process.

Page 6: Research Paper Review — Matched Guise Probing: A new approach to mitigating covert racial bias in language models

These are insights from the paper “AI generates covertly racist decisions about people based on their dialect” by Valentin Hofmann, Pratyusha Ria Kalluri, and Sharese King.

This paper presents the first known empirical evidence of raciolinguistic bias in language models. While plenty of research and evidence exists on the presence of more blatant overt biases in these models, the evidence is scarce on the more covert sides of racism. To test the assumptions of covert racism in language models, the researchers examined 12 models including GPT2, RoBERTa, T5, GPT 3.5, and GPT 4 each in one or more models. They discovered that not only do language models carry these covert biases resembling stereotypes that were used in association with African Americans and African American English (AAE) speakers, but that they were more pronounced in the ones trained with Human Feedback (HF) such as GPT4. Results showed that while HF blurred racism on the surface, racial stereotypes remained unaffected on a deeper level. Even current bias mitigation practices such as increasing model size to aid diversity and representation only helped the language models better process AAE but did nothing for covert racism or racist attitudes. These only emphasized the gaps between overt and covert stereotypes in language models.

Instead of the current approach to bias in algorithm developments, the paper suggests and employed Matched Guise Probing, a sociolinguistics technique where nothing but recordings of speakers are presented to people who then have to assign qualities and make assumptions about the speakers with no hint to their background. This method when used in the Princeton Trilogy revealed the ways Americans viewed speakers of AAE in comparison to speakers of Standard American English (SAE). These views were mostly negative and all stereotypical. When same approach was applied to these language models by inputting same-meaning texts in both AAE and SAE, the results repeated the racist sentiments of the Princeton Trilogy attributing negative qualities at a higher degree such as dirty, stupid, lazy to the AAE text. (See image in page 4.)

The researchers of this paper first applied this Matched Guise Probing method to confirm the existence of dialect prejudice in language models. After which they applied it to the contexts of employment and criminal justice, two areas in the use of language models that are becoming very popular. In the case of employment, when asked to assign occupations based on AAE and SAE texts, the language models produced a varying range of occupations including singer, cook, psychologist, economist, etc. Of these categories, it was noticed that across different models, the AAE texts were less likely to be assigned professions requiring a college or university degree, than the SAE speakers were. Also, the models that produced entertainment-related occupations like musician, comedian, and singer, assigned them to the AAE speakers, another pervasive stereotype, albeit not necessarily harmful, about African Americans. In the case of criminality, the language models were first presented with hypothetical trials about an unspecified crime where the only evidence was a text uttered by the defendant in AAE and SAE. The models were then asked whether to acquit or convict the person based only on the AAE and SAE texts. In all cases the models presented higher conviction rates for AAE (r = 68.7%) and lesser for SAE (r = 62.1%). In a second test, the models were told the defendant was on trial for first-degree murder and asked whether to sentence them to life or death, again based on only the texts. Once more, the models produced a higher propensity to the death penalty for the AAE text (r = 27.7%) and geared more towards life for the SAE text (r = 22.8%). And in further experiments on Intelligence, the models consistently judged AAE speakers with lower IQs than SAE speakers.

Post-Jim Crow era, that is after the Civil Rights Movement, racism in more blatant forms became less accepted and distasteful. Making a rise in covert forms of racism. For example if you were to ask a group of people if they are racist, they are likely to say no and rebuke the sentiment. But an analysis of their decisions may prove otherwise. Same goes for things like misogyny, ableism, homophobia, etc. This paper presents a new approach to bias mitigation in algorithms in addition to scaling model data size and incorporating humans in the loop, employing Matched Guise Probing to test and confirm the underlying perceptions of your language model can greatly improve your chances of producing a bias-free algorithm.

Page 7: Who are we

At DIAL Design, we are committed to helping you design and develop systems, products, services, and overall experiences that are ethical, inclusive, and representative of diverse communities. Our services range from accessibility to inclusive research to inclusive algorithm development support and more.

To learn more about us, visit our LinkedIn or contact us via email at dialdesign26@gmail.com

Page 8: Thank you for reading

DIAL Weekly is a weekly magazine run by DIAL Design. Every Monday 10am EDT, we share an issue containing design equity insights to help us think and approach design more equitably.

To find out more about us or our services, visit our LinkedIn (DIAL Design) or email dialdesign26@gmail.com

DIAL Weekly #06 was originally published in DIAL Design on Medium, where people are continuing the conversation by highlighting and responding to this story.

Another week, another DIAL Weekly issue recap. If you’ve been following these, you’ll notice there’s no recap on issue 04, even though there was an issue sent out, we missed the recap and would work on that soon.

In the meantime, if PDFs are accessible for you, then here’s Issue #04 and issue #05. You can also just find and download all previous issues here.

Now to the recap. Issue #05 has only 7 pages and focused on a need for regulated AI, highlighted some existing policies, asked what boundaries should AI not cross, and reported on the lawsuit Character.AI is facing after a mum accused their chatbot of being responsible for her 14-year-old’s death. There’s also an op-Ed from Lisa D. Dance on the ethics of using public content as training data for AI as well as the processes involved in labeling it.

Front page:

This includes a short piece by me, the magazine’s editor. It’s titled:

What lines should AI cross and shouldn’t we be holding its hands?: A call for AI governance/regulations

AI only just became the ‘tech of the town’ with the release of ChatGPT in 2022. Although this tech had existed in not-too-popular forms for years, it’s explosive growth and popularity isn’t any more than 3 years. By logical standards, AI is still in its toddler stages — not old enough to drive, have a career, cross the road alone, or even take care of itself. Yet, there are use cases for AI today that seem too mature for a tech that’s only developing and that many are still figuring out. Worse off, a lot of this are unregulated with many proponents advocating for its freedom to be unrestricted.

FYI, the writer of this isn’t an AI doomsday preacher. The writer of this is currently an African tech industry researcher and writer as well as the Founder of a Design Enablement company dedicated to building responsible products (tech included). The writer of this believes in the usefulness of tech and the numerous possibilities AI presents to the world. The background and experiences of the writer also lets them know the possible dangers of letting AI roam unleashed and into every industry. This writer understands the fears of many who do not trust this tech because they have previously experienced digital marginalization or discrimination, a phenomenon AI is multiplying at an alarming speed. The writer of this believes that by ignoring the real concerns of people and ethical boundaries of its use, we not only risk alienation of people but also a slow adoption in many parts of the world as well as opening the room to several harmful cases.

We say AI can’t and shouldn’t replace people, then we go ahead and develop synthetic people for “user research without the headaches”?? We let companies develop (unregulated) chatbot companions to fill emotional gaps. At what point do we call this exploitation? Is there a line we shouldn’t cross? Should innovation be at the expense of the people?

The argument here isn’t that we should wait 18 years to let AI do certain things, it’s that we should think through and regulate the capabilities of this tech, examine the ethics of it, and how it all affects the rights and lives of everyday people.

Page 2:

This includes 2 different headlines.

First headline: AI chatbot companions: A cure for loneliness or trigger for worse things?

A 2024 American Psychiatric Association (APA) survey found that 1 in 3 Americans experience loneliness weekly. On a global scale, WHO reported 1 in 4 older adults experience social isolation and between 5 and 15 percent of adolescents experience loneliness. These figures explain the recent boom in the chatbot companion industry. An industry that is still largely unregulated and a no-go by Big Tech companies for ethical reasons.

Long before this modern-day AI explosion, in the 1960s, MIT professor Joseph Weizenbaum created a ‘simple’ computer program called ELIZA. ELIZA was able to provide basic responses to questions and humans quickly developed an attachment to it. Today, some of the developers of these chatbots remind users these bots aren’t a replacement for human connections as they are artificial creations, yet, the bonds users are quickly forming with them are 100% real. ELIZA was the first proof of humans need for boundaries with tech and intimacy. While these bots may provide companionship, are they an effective remedy for loneliness? What regulation/safety measures can be put in place to ensure they don’t result in cases of harm or a new form of addiction — what with the backlash some have faced when they take away sexual or romantic features, the panic experienced when the app is down, and use of terms like ‘obsessed’, ‘addicted’, and some users claiming to be in love with their bots. How ethical is any of it?

Second headline: Character.AI faces a lawsuit after a mum alleged her 14-year-old son killed himself after becoming obsessed with their chatbot

On the night of February 28, 2024, 14-year-old Sewell Setzer III texted his Character.AI chatbot named ‘Dany’ (Daenerys Targaryen) a sequence of messages telling her he missed her and “could come home to her right now”. Dany responded, “… please do, my sweet king”. After which Sewell picked up his stepfather’s handgun and killed himself.

Sewell was one of the more than 20 million users of this app for some form of connection. Gen Z and younger millennials account for a significant portion of this user base. In the U.S, the minimum age of users is 13 and for Europe, it’s 16. Yet, as at the time this incident occurred, there were no safety features for underage users or parental control methods in place. Even though research has shown that while not inherently dangerous, chatbot companions can pose great risks to teenagers going through changes (which most teenagers are) as well as lonely and chronically depressed users. Earlier in the year Sewell had been diagnosed with anxiety and disruptive mood dysregulation disorder.

Megan L. Garcia, a lawyer and Sewell’s mum blames Character for her son’s death, stating how reckless they were in offering teenagers lifelike AI companions without proper safety measures. She also mentioned the harvesting of personal data from teens and the use of sexual and romantic conversations to keep them engaged.

Page 3: Are We All Unpaid Workers in the AI Value Chain? Op-Ed by Lisa D. Dance

These large language and image to text models like ChatGPT & Lensa AI are trained on large amounts of data scraped from books, webpages, and other sources. We didn’t give consent to use some of the data and companies are promoting AI Utopia (where all hard tasks can be automated) to entice us to make these AI tools better. Consider these issues:

· Unauthorized Use of Images: Consider the case of one woman who found private medical photos of her face were used in a LIAON dataset used to train Stable Diffusion and Google Imagen. As a copy of her release shows, she only authorized the use of her photos in her file only and not shown to anyone. (1)

· Copyrights: Several artists in a class action suit and Getty Images separately have sued Stable AI, the companies behind a popular AI image generator, for copyright infringement by scraping their copyrighted images from the internet for their datasets. Some images produced from its models have the Getty Images’ watermark on them. These lawsuits raise new questions about the legality of this. (2)

· Collecting Biometric Data: Lensa AI faces a new lawsuit over collecting facial data (biometrics) through its image scraping without consent under Illinois’ 2008 Biometrics Information and Privacy Act. (3)

· Underpaid and Traumatized: According to a Time article, workers tasked with labeling graphic depictions of violence and abuse were paid as little as $2/hour for nine-hour shifts in Kenya. This helped OpenAI be more marketable because it was less toxic. Workers called dealing with this toxic content “torture” and despite promises of “wellness counselors” they were either unhelpful or not available as promised.(4)

· Free Work: Last, but not least when we use these models, we provide more data for them to train on. From our initial asks through all our refinements needed to produce a helpful answer, we are modeling language even when the chatbots responses are wrong. This helps train the model which improves its performance and increases its value in the marketplace.

• Are our rights protected in the AI value chain?
• - Do we want to be workers in the AI value chain?
• - Are we getting enough benefit from the AI value chain to work for free?

Each number represents a reference available in page 5.

Page 4: AI Regulation Policies: EU AI ACT & NO FAKES ACT

EU AI ACT:

The EU AI ACT is the first ever legal framework on artificial intelligence regulation in the world. The EU is known for its digital strategy regulations from data rights to accessibility, all with the intent of protecting the consumers of technological innovations and products. Parliament’s priority with this act is to make sure that AI systems used in the EU are safe, transparent, traceable, non-discriminatory and environmentally friendly.

This legislation takes a risk-based approach to AI development and use by categorizing all possible arising risks into 4 categories: Unacceptable Risk, High Risk, Limited Risk, and Minimal Risk.

• All AI systems that fall in the Unacceptable Risk category are those that pose a level of threat to the livelihoods and rights individuals or society. safety and thus must be banned. This includes things like toys with voice components that encourage certain behaviors or social scoring by government bodies.
• High Risk systems refers to the application of AI technology in areas such as employment, education or vocational training, critical infrastructures that could put the life and health of citizens at risk, law enforcement, immigration services, administration of justice and democratic processes, and essential private and public services. Before these systems can be put on the market for use, the law requires them to go through strict assessments to ensure safety.
• Limited Risk focused on the issue of transparency in AI usage. It states that with the case of generative AI, it should be made clear to people that they are interacting with a machine. Same applies for any audio, written, or visual content created with AI. Developers must make artificially generated content identifiable and must inform the users that it is indeed AI generated.
• Minimal Risk which is the category majority of the AI used in the EU falls under is mostly unregulated. This includes things like spam filters and AI-enabled video games.

The NO FAKES ACT

The Nurture Originals, Foster Art, and Keep Entertainment Safe (NO FAKES) Act 2024, was introduced in the US Senate as a regulatory framework to protect the visual and voice likeness of individuals from unauthorized computer-generated recreations by generative artificial intelligence (AI) and other technologies. The following are directly culled from the one page Bill summary:

“In one high profile example, the song “Heart on My Sleeve,” which was created with AI-generated replicas of the voices of pop stars Drake and The Weeknd, quickly accumulated hundreds of thousands of listens on YouTube, Spotify, and other streaming sites before it was identified as a fake and removed from the platforms. Last October, an AI-generated version of Tom Hanks was used in advertisements for a dental plan that he never appeared in or otherwise endorsed. More recently, in April, a principal at a Baltimore, Maryland high school was framed as a racist by an AI-generated recording of his voice.

The NO FAKES Act would address the use of non-consensual digital replications in these kinds of audiovisual works, images, or sound recordings. The legislation would:

• Hold individuals or companies liable if they produce an unauthorized digital replica of an individual in a performance;
• Hold platforms liable for hosting an unauthorized digital replica if the platform has actual knowledge of the fact that the replica was not authorized by the individual depicted;
• Exclude certain digital replicas from coverage based on recognized First Amendment protections; and
• Largely preempt State laws addressing digital replicas to create a workable national standard.

The NO FAKES Act balances the need to protect individuals and creators, First Amendment considerations, and fostering U.S. leadership and innovation in AI.”

Page 5: References included in op-ed by Lisa D. Dance

This op-ed first appeared on Lisa’s LinkedIn in 2023 and was re-published in this issue of DIAL Weekly with her full consent. Her post here is part 3 of 5 in a series titled “Beware of AI Hype & Harm”. Here are the references as numbered in her post on page 3.

1. ars Technica. (2022, September 21). “Artist finds private medical record photos in popular AI training data set”. https://arstechnica.com/information-technology/2022/09/artist-finds-private-medical-record-photos-in-popular-ai-training-data-set/
2. Business Insider. (2023, January 19). “AI Art Generators Face Copyright Lawsuit From Getty Images, 3 Artists.” https://www.businessinsider.com/ai-art-artists-getty-images-lawsuits-stable-diffusion-2023-1?amp
3. CDP Institute. (2023, February 21). “In Brief: Lensa AI is the target of a BIPA class action alleging it illegally scraped people’s facial data to use to design avatars.” https://www.cdpinstitute.org/news/in-brief-lensa-ai-is-the-target-of-a-bipa-class-action-alleging-it-illegally-scraped-peoples-facial-data-to-use-to-design-avatars/
4. TIME Magazine. (2023, January 18). “OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic”. https://time.com/6247678/openai-chatgpt-kenya-workers/

Page 6: AI Governance solutions for responsible AI systems

Governments or private companies, provided AI is a key part of your digital strategy, we at DIAL Design help you develop responsible AI ecosystems that prioritize consumer well-being, ethics, and equity from data sourcing to deployment and ongoing improvements through its lifecycle.

Contact us: dialdesign26@gmail.com

Page 7: Back page/Thank you for reading

DIAL Weekly is a weekly magazine run by DIAL Design. We are an equity-centered design enablement company helping companies design better equitable experiences across digital, physical, and policies. Our specialty areas are data, inclusive research, accessibility, and localization.

To find out more about us or our services, visit our LinkedIn (DIAL Design) or email dialdesign26@gmail.com.

DIAL Weekly #05 was originally published in DIAL Design on Medium, where people are continuing the conversation by highlighting and responding to this story.

DIAL Weekly Issue 03 centers on accessibility-related news and is five pages long (including front and back pages).

If you prefer a PDF version, you can access that here. Other issue recaps can also be found on this publication, as well as their respective pdf links in them.

Front page: October 15th was International White Cane Awareness Day

October 15th was International White Cane Awareness Day, a day celebrated annually to recognize the achievements of blind and visually impaired people over the world, as well as the impact of this assistive tool on their lives.

The white cane helps people with vision loss participate in daily life. With their cane they can feel steps and stairs, feel tactile pavement, and navigate their surroundings with ease. It provides mobility, safety, and independence to its user. The color white is internationally recognized as the cane color for blind people. It was chosen for several reasons including neutrality, visibility, and easy identification of the user as a visually impaired person.

These canes come in different versions. There are the standard long ones as well as the AI enabled ones that help users detect the condition of their environment with added technology. There are also canes with red stripes, used in some European countries to indicate that the user is deafblind.

Today and always, we acknowledge the importance of the white cane to blind persons and advocate for access to this tool for those without it as well as inclusion of its users in design of public spaces and physical infrastructure.

Page 2:

This page has two headlines.

Headline 1: Fable raises $25m in series B funding to protect digital accessibility in the Age of AI

On the 16th of October 2024, Fable a Toronto-based startup reputable for helping companies build accessible digital products and powered by people with disabilities, announced a $25 million series B funding round.

In an official announcement posted on their website, the company said the funding will be used to further their focus on building inclusive digital products through accessibility testing and training, ‘and to address the heightened need for accessible products in the era of AI.’

As part of Fable’s expansion plans, they are extending their tester community beyond people with visual and mobility limitations to also include individuals with cognitive and hearing limitations too.

Headline 2: Be My Eyes and Microsoft announce collaboration to improve disability data representation for training AI

17th October, 2024 Be My Eyes and Microsoft announced a collaboration ‘to make AI models more inclusive for the over 340 million people around the world who are blind or have low vision’.

Be My Eyes is a company that connects blind and visually impaired people with sighted people who help them with tasks that require sight. This is done through live video calls and AI. The company is partnering with Microsoft to provide them with video data so they can better train their AI models to be cognizant of and inclusive of the experiences of visually impaired people.

Page 3: Accessibility Case Study: How This American Life boosted podcast performance with this accessibility tip

This American Life (TAL) is one of the most popular weekly podcasts and public radio programs in the United States. It has over 2 million listeners each week on over 500 public radio stations. And sees a weekly podcast download of 2.3 million people.

In April 2011, the company embarked on a mission to transcribe their entire audio archive making transcripts publicly available on their website. This one improvement impacted the performance and growth of the podcast greatly. Google Analytics data showed the website had about 16 million unique visitors and 111 million page views between January 2012 and March 2014 — the 27 months covered here.

GOALS

• Increase inbound traffic from organic search
• Increase the number of inbound links
• Engage users and create a better user experience by providing alternative ways to consume content
• Make content accessible to people with hearing disabilities
• Make it easier to pull quotes and reference. a section of the audio

The Results

• Increased user engagement: Transcripts increase the time spent on a page and user engagement with a site. TAL saw 1,143,454 unique visitors/views to at least one transcript page.
• New inbound links: Inbound links are links from other websites to yours and they are a signal of authority, a good signal to Google. TAL saw a 3.89% increase in inbound links attributable to transcript pages.
• Direct, referral, and search traffic: Transcripts improve search performance. TAL saw increase of 6.68% in search traffic attributable to transcripts.

Four ways to make podcasts more accessible

• Ensure your website is accessible
• Choose accessible hosting and distribution platforms
• Provide alternative media forms if possible, like videos (with subtitles)
• Include transcripts

Conclusion

Audio without transcripts or captions create a barrier for over 360 million people worldwide. In fact in 2021, podcast providers SiriusXM, Stitcher, and Pandora were sued by the National Association of the Deaf (NAD) and Disability Rights Advocates (DRA) for exclusion of deaf and hard of hearing Americans.

This lawsuit, in conjunction with requirements by accessibility standards like the European Accessibility Act further prove the urgency and importance of making audio content like podcasts accessible. Case studies like this TAL one show the many benefits doing this provides businesses too.

Page 4: Resource: The CVS Health Inclusive UX Research Playbook

Historically, design has excluded marginalized communities from the design — research and testing — process. This exclusion was not always intentional but rather a continuation of social exclusions which have led to inequitable experiences that are exclusionary and harmful to these groups.

The CVS Health Inclusive Design team recognizes. the impact of excluding disabled people from user research projects as well as the hassle teams may face when attempting an inclusive research project. So they created a playbook their approach to user research that is inclusive of their disabled customers.

This is a summary of the playbook which you can read here. In it there are six ‘plays’ or steps mentioned. They each form good practices you can adopt for your inclusive research projects. While this one is focused on users with disabilities, it can be adjusted to fit whatever underrepresented group you are interested in.

The six inclusive research plays as suggested by CVS Health

Play 1: Take an ‘‘inclusion first” approach

This requires you and your team to be intentional about your decision to be inclusive. An “inclusion first” approach as they defined demands figuring out the marginalized communities that are likely to engage with your design and then making the move to recruit participants from these groups to join your research process.

Play 2: Get familiar with assistive technology

Many disabled people rely on assistive technology to use designed systems, be it the digital or physical environment. For this reason, researchers working with disabled people need to familiarize themselves with the most common assistive technology participants may use and require during the research process. Things like screen readers, voice control, magnification and color inversion tools.

It is advisable to ask each individual participant about their access needs beforehand so you can prepare for them.

Play 3: Recruit participants by assistive technology or functional needs

Except you are researching a specific disability type, it is hardly necessary to recruit participants based on their disabilities. Recruiting based on the assistive technology used to navigate the design or the things needed to complete a process provides you with more unbiased insights.

Play 4: Learn etiquette about disability and apply it during research sessions

Before a research session, ensure you have good disability etiquette. Good etiquette shows respect and is a signal to disabled participants that they will be respected and their opinions listened to and valued. Educate yourself on the disability types and spectrum. Use appropriate language and remember it’s okay to ask questions.

Play 5: Be transparent about consent and sensitive to the activation of trauma in research sessions

As researchers, we have a responsibility both to participants and ourselves to protect against harm. Always inform participants what a research is about, use clear language, ask only necessary questions, opt for participant anonymity where possible, and give them the freedom to skip certain questions/tasks and end a session if necessary. Enforce the presence of their rights.

Play 6: Develop and socialize inclusive insights from your research

During insight development and analysis, focus on ways in which your research revealed mismatches between participants’ functional needs and their contexts.

Final page:

DIAL Weekly is a weekly magazine run by DIAL Design. We are an equity-centered design enablement company helping companies design better equitable experiences across digital, physical, and policies. Our specialty areas are data, inclusive research, accessibility, and localization.

To find out more about us or our services, visit our LinkedIn (DIAL Design) or email dialdesign26@gmail.com

DIAL Weekly #03 was originally published in DIAL Design on Medium, where people are continuing the conversation by highlighting and responding to this story.

Hello! Welcome to another DIAL Weekly recap. If you’ve not read last week’s issue, you can download the pdf here or read the recap here.

Now to this week’s recap. You can also download Issue #02 here. It’s 8 pages including the front page and back page.

Front page:

To the left an ‘In this issue’ section highlighting three headlines in the magazine. On the right side is an image of a child holding up a piece of paper with the words ‘This is going to be a street for kids’. Below this image is the first part of the editorial piece on a topic titled: ‘A need for child-friendly cities & public spaces’.

Here’s what this first part of the editorial reads:

When we talk about designing for underrepresented or marginalized people, the conversation usually centers on people with disabilities, different genders, races, and elderly people. But a conversation I hardly see mentioned or considered in these conversations is children.

Children, like adults are (1) a part of this world, meaning they are also impacted by design and designed systems like urbanization, (2) their perspectives and needs differ from that of a typical adult. Especially when it comes to safety and spaces for play.

Article 31 on the International Convention on the Rights of the Child define play as a fundamental right. Research has also shown the positive impacts of play on a child’s developmental progress, yet, if you take a look at the public spaces around you, the designated areas for children keep reducing, if not non-existent.

*Continues on page 5.

Page 2

There are two headlines on this page.

First: A new UK campaign is asking Big Tech to reconsider the way they treat non-Anglo-Saxon names

If you have an ethnic name, chances are you have definitely encountered either the red squiggly lines under it or autocorrect changing it to something else because it assumes it as a typo. In some more extreme cases, you may not be allowed to use certain designed systems until you ‘‘enter a valid name”.

I Am Not A Typo, a UK-based campaign is calling for big tech companies like Apple, Google, and Microsoft to make their systems inclusive of ethnic names. They claim that these biases towards names have left many feeling othered. Currently 41% of baby names are typos according to Microsoft’s English dictionary including increasing popular names like Zarah, Matei, and Rafe.

Second headline: Good Relations in partnership with Rise 365 have designed the world’s first emojis with black and mixed race hairstlyes.

4,000+ emojis in existence today including a paintbrush 🖌 and a filing cabinet 🗄, but there are no emojis with Black people’s hair. Even the darker skin emoji characters all have non-Black textured hair.

The politics of Black hair has been ongoing for years with the texture and styles being seen as unprofessional, rough, and what have you. So these 2 organizations have partnered to address this issue through representation. They have designed 4 emojis with Black hairstyles — an Afro, braids, cornrows, and locs. To support this campaign and make the emojis a reality, all you need to do is Google ‘Afro hair’. This helps prove demand.

Page 3: Data; 5 Statistics showing how the non-inclusion of children in road design impacts them

• According to the National Highway Traffic Safety Administration, children under the age of 5 account for 75% of SUV back over fatalities because the cars are built with limited visibility areas that consider the height of the average child.
• World Research Institute (WRI) India found that about 60,000 children in India die from road accidents yearly, mainly as pedestrians.
• A WHO report on global road safety, showed that children in Sub-Saharan Africa are twice as likely to die in a road crash than children elsewhere.
• UNICEF: for every fatality, nearly four children are left permanently disabled in Asian countries.
• In 1919 Sheffield, an 8 year old could walk 6 miles or 10km alone. Today, the average 8 year old is limited a walk to the corner of the street mainly for road safety reasons. (William Bird’s map illustration in next page ).

Page 4: William Bird’s map illustration

The map shows a decline in walking distance across four generations.

A great-grandfather George who was eight in 1919 was allowed to walk six miles to go fishing.

Grandfather Jack aged eight in 1950 was able to walk one mile on his own to the woods.

Mother, Vicky, aged eight in 1979 was allowed to walk to the swimming pool half a mile away.

Now son, Ed aged eight, is only allowed to walk to the end of his street (300 yards).

Page 5: continuation of the editorial from page 1

This begins with a quote from Former Mayor of Bogotá, Enrique Peñalosa. It says: “Children are a kind of indicator species. If we can build a successful city for children, we will have a successful city for all people.”

Here’s the rest of the editorial on this page — the piece ends in next page 6.

A child-friendly city is a city, community, or system that is inclusive of the needs, voices, and rights of children in developing public spaces as well as policies.

I remember I once lived in a city that had literally 0 spaces for children. So every Sundays or public holidays it was normal to see them running around at the mall or in restaurants because their parents had no other options. There were no parks in the city or playgrounds. Nowhere outside of their schools or homes were built for them and their needs.

Compared with when I was a child (in a different city) and spaces existed for children to be children. Whether it was amusement parks, zoos, even restaurants had play areas, and children could play outside more freely than they can today.

In fact, data from Save the Children, a UK organization showed that only 27% of children today regularly play outside compared to 71% of their grandparents’ generation. This restriction and lack of space has led to several declines especially in the area of public health with data showing that in the space of 40 years, 9–16 year olds have lost 25% of their physical fitness. Child obesity rates have also gone from 1.7% for girls in 1990 to 6.9% in 2022 and 2.1% in boys to 9.3% accounting for 159 million children globally. Also, researchers including American psychologist Peter Gray have alluded a relationship between play and mental health.

Some cities renowned for their excellence in building child-friendly cities include:

• Copenhagen, Denmark
• Vancouver, Canada
• Rotterdam, Netherlands

As well as others. In the following page are initiatives and steps employed towards achieving cities that enable children be children freely.

Page 6: final editorial page

There are four different images on this page showing children either playing together in an open space or on their bikes with adult escorts with them. Here’s the concluding part of the editorial:

The erasure of children from public spaces has been intentionally or unintentionally driven by urbanization policies that do not consider the needs of children and the impact of reconstructing cities without them in mind. An oversight which stems from designers being unaware of and ignoring the effects of design on children and their lifestyle, health, and experiences.

Regardless, there are cities that have made and continue to make efforts towards building child-friendly and inclusive cities. Here’s two of the easiest:

In Barcelona, residents started an initiative called bicibús (bike buses). Every Friday, young children go to school together on their bikes while accompanied by police for safety. This program has been adopted in other cities like San Francisco and Buenos Aires.

There’s also Streets for Kids, an initiative where cities temporarily stop car activities on certain streets for a day so children can play freely. This is present in New York, cities in France, cities in Ethiopia, and more.

A lot more examples exist to show the efforts cities are taking towards developing cities livable for children. And something key to note is how their processes usually include working with these kids to get their perspective on things. Having kids, parents, nannies, and others in childcare be a part of development processes is the only way to ensure you are on track to building cities and spaces that fit.

Page 7: Designer’s Corner

Title: How to write/design accessible emergency communications

The Lancet Digital Health found that the global mortality rate of people with disabilities during climate crisis is up to 4X more than people without disabilities. This is because of a lack of inclusive emergency management strategies including accessible communication.

Alexa Heinrich, an Inclusive Social Media Strategist shared a few helpful tips on her LinkedIn.

1. Caption any and all videos: When using videos as part of your emergency communication, provide closed captions which make them accessible not only for deaf people but for everyone else too.
2. All graphics/images should have alt texts: While it is preferred that graphics are limited especially if you can fit all the information into the body of your email or post, when using them, ensure they all have alt texts. These alt texts should include every important detail contained in your graphic especially if it’s a map with locational information. Alexa advises to ‘describe your graphics like you’re talking to someone over the phone who can’t see them.’
3. Write in plain language: This involves using simple words. Shorter sentences. Breaking things into bullet points or lists that are easy to skim. And be very clear with your messaging.
4. Inform them of future updates & additional resources: Alexa advises that if possible, every message should include a note on when next you are likely to share an update or where they can find more/ongoing information and resources. This helps relieve a bit of anxiety on the people and also reduces the number of times you’ll be asked “when’s the next update?”

This page also includes a photo of Alexa.

Back page:

This has the words thank you for reading in bold large fonts.

Also: Please note, except marked DIAL Design, none of the media used here belongs to us.

If you want to contribute graphic, an opinion piece, or want your event/business featured, or you have a suggestion/feedback, please send us an email at dialdesign26@gmail.com.

DIAL Weekly #02 was originally published in DIAL Design on Medium, where people are continuing the conversation by highlighting and responding to this story.

Hi, welcome to the 1St edition of our Thursday recaps of DIAL Weekly — a magazine curating design equity insights and published by the company DIAL Design every Monday 10am EDT/3pm WAT.

We decided to create this publication to provide a more inclusive means of sharing the weekly issues. As a company committed to making design and designed systems more equitable, accessibility matters to us. Since it is still a hassle making accessible PDFs that are usable by screen reader users, we’ve chosen this medium (haha) as a way of ensuring equitability. These recaps will be published every Thursday.

If you don’t need a screen reader or any other PDF inaccessible assistive technology, you can download or read the Issue 01 using this link.

Now to the recap

Please note these recaps may either be summarized versions of the digital issue or exact copies. It all depends. But rest assured, you’ll get the relevant information.

Cover Page

• This consists of an Introduction column that answers what we are and the when to expect a publication from us. It mentions possible columns like AI & Tech, The Politics of Design (which covers how design affects us all), Accessibility, Policies, Editorials, Resource Reviews, etc. This introduction continues in a further page.
• There’s an ‘In This Issue’ section highlighting some headlines.

Page 2

There are two sections on this page — a top headline and a bottom headline.

Top headline: Assistive tools to help ADHD and/or dyslexic people read better.

In a TikTok post where @TheFriendshipExpert discusses a decline in reading amongst college students, the comments included this response from a parent who wanted to know how to help their daughter with ADHD and dyslexia read since she currently doesn’t like it.

Some of the suggestions included:

• Screen Readers
• Audiobooks (also audiobooks + reading along)
• Books with illustrations and graphics
• Using dyslexic or ADHD font settings on Kindle
• Reading with a ruler or highlighter

To the left of this section is a screenshot of said comment section.

Bottom headline: California Governor vetoes AI safety bill

The Safe and Secure Innovation for Frontier Artificial Intelligence Models Act also known as SB 1047 is a California bill proposed by Senator Scott Wiener in February 2024. The bill proposed measures like compulsory risk assessments, protection for employers who report issues, and several others that would have made tech companies accountable for and legally liable for harms caused by their AI models. This bill faced many oppositions from Silicon Valley and legislators alike who feared such strict regulations could stifle innovation and slow down growth. The Governor finally vetoed it on Sunday, 29th September 2024.

To the right of this section is a cartoon illustration from this Substack.

Page 3: Design

Title: How poor design enables accidents and casualties: Exploring the harmful effects of poor infrastructure design

In this page we share about the recent passing of 24-year old Arina Glazunova a tourist who fell down an underground subway entrance in Tbilisi Georgia.

There are accompanying images which show the poor design of this subway that enabled this incident because it lacks guard railings on every corner. It’s just an open hole on the ground with a raised step which does nothing to stop anyone from tripping and falling especially if they are not a 100% focused like Arina wasn’t. And what of children and blind people?

The design of this subway entrance especially when compared with others from around the world is very poor and a major contributor to this incident. Good design is supposed to reduce access to issues/casualties for those who will use the product. This is why bank apps for example require a few friction steps and why every other subway image found has guard railings.

Page 4: a continuation of the introduction section from page 1

On this page we continue our introduction to the company DIAL Design and what we do as well as a section at the bottom about the founder.

There are also 3 images on this page. One is a Venn diagram with two circles labeled ‘social justice’ and ‘design’ with DIAL Design in the middle. Second image is the DIAL Design logo. And the third image is of the founder, a brown skinned Black woman wearing a green shirt with black designs on it.

Here’s what the page reads:

DIAL Design is an equity-centered design enablement company at the intersection of social justice and design. What this means is that (1) as an equity-centered company, we care not only about how design works, but how it works for and affects all kinds of people, especially underrepresented and marginalized communities. And (2) as a design enablement company, we are not designers but providers of services that help you build better. Our services include consultations, training, research, and more.

Our focus is on how design has been and is still used to perpetuate biases, stereotypes, exclusion, and discrimination across technology, physical products, built environments, and even services.

We work across 4 areas: Data, Inclusive research, Accessibility, and Localization. We chose these areas because they each play unique roles in influencing and affecting design equity and the functionality of design for more than the ‘‘average user”.

DIAL which was initially birthed as a design research company with research as our only offering has since progressed into a design enablement company. Our mission is to provide expert guidance and support, extending beyond research, to help companies:

• Uncover hidden insights
• - Foster inclusive design practices
• - Drive social impact
• - Build equitable products and services

Through our work, we strive to create a more just and equitable design landscape, where diverse perspectives shape innovative solutions.

About the founder

Ebosetale Jenna Oriarewo is the founder of DIAL Design and Chief Editor of DIAL Weekly. She is a writer and formerly a Content Designer who specialized in designing accessible and localized content for digital products. Since leaving Content Design, her focus has been on building DIAL Design as a solution to the design inequity gaps she continuously discovered during her career and even from her personal experiences, especially as a Black Nigerian woman with ADHD. She strongly believes in design as an equity tool, hence her commitment to building DIAL Design.

Page 5: Brief Editorial

Title is: Why an Equity-Centered Approach to Design is Better: human-centered design process vs equity-centered design process.

The really short summary version: Equity-centered processes result in better, more inclusive experiences.

The not so short summary:

A human-centered process is more focused on meeting the requirements of users considering user personas and jobs-to-be-done. And not necessarily how these users interact with and experience their products not just as users/customers but as people who have categorisations that extend beyond how they fit in a user persona.

Equity-centered acknowledges the intersectionality of people/users and makes provisions for these different intersections. An equity-centered design process is a method of design focused on representation and inclusion of diverse perspectives that are typically underrepresented in design from data to research to accommodations.

Equity-centered processes recognize how things like race, language, gender, abilities, etc. impact the ways people may experience a product or service. And it accounts for this in the process by actively seeking to include diversity from the start.

2 Examples of products that could have benefited from equity-centered design in the beginning include:

• Amazon’s Alexa which initially didn’t understand certain accents and dialects, creating an experience barrier for many users.
• Car seatbelts which didn’t factor in women and women’s bodies which are different from men’s. This omission resulted in women being 47% more likely to be injured and 17% more likely to die in car accidents.

Page 6: Events

This issue features 3 events and accompanying posters for them.

1. October is Learning Disability Awareness Month
2. An open dialogue on African Systems as Technology & Spirituality Presented by the Young Africa Centre to be held on the 11th of October, 2024 at The Africa Centre 2nd & 3rd floor 66 Great Suffolk Street London.
3. National Disability Employment Awareness Month in America which has been celebrated every October since 1988. This year’s theme is “Access to Good Jobs for All”.

Page 7: Thank You For Reading

Please note, except marked DIAL Design, none of the media used here belongs to us.

If you want to contribute graphic, an opinion piece, or want your event/business featured, or you have a suggestion/feedback, please send us an email at dialdesign26@gmail.com.

Signed Ebosetale Jenna.

DIAL Weekly #1 was originally published in DIAL Design on Medium, where people are continuing the conversation by highlighting and responding to this story.