<?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 Amanda Pitcher on Medium]]></title>
        <description><![CDATA[Stories by Amanda Pitcher on Medium]]></description>
        <link>https://medium.com/@amanda-r-pitcher?source=rss-2d1924ebe000------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*r54QnJViVhBVcGzJ</url>
            <title>Stories by Amanda Pitcher on Medium</title>
            <link>https://medium.com/@amanda-r-pitcher?source=rss-2d1924ebe000------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 23:19:14 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@amanda-r-pitcher/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[Meet the Brilliant Jerk Hiding in Your AI]]></title>
            <link>https://medium.com/versent-tech-blog/meet-the-brilliant-jerk-hiding-in-your-ai-ef029ed06fdd?source=rss-2d1924ebe000------2</link>
            <guid isPermaLink="false">https://medium.com/p/ef029ed06fdd</guid>
            <category><![CDATA[ethical-ai]]></category>
            <category><![CDATA[ethical-ai-development]]></category>
            <category><![CDATA[persona-testing]]></category>
            <category><![CDATA[compassionate-design]]></category>
            <dc:creator><![CDATA[Amanda Pitcher]]></dc:creator>
            <pubDate>Sat, 07 Mar 2026 03:35:43 GMT</pubDate>
            <atom:updated>2026-03-07T03:35:43.412Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Syrena Rath’s Three Frameworks for defeating the biased villain in your AI." src="https://cdn-images-1.medium.com/max/1024/1*-9VhuXololAmx5M7jAWDMQ.jpeg" /><figcaption>Syrena Rath’s three frameworks for defeating the biased villain in your AI, from the keynote at DDD Melbourne 21 February, 2026.</figcaption></figure><h4>ethical AI</h4><h4>AI doesn’t have bad intentions. It doesn’t need to. Here’s why that’s the problem.</h4><p>AI is already making decisions about who gets hired, who gets healthcare, and in the most devastating cases — who lives. At <a href="https://www.linkedin.com/company/dddmelbourne/">DDD Melbourne</a>, Saturday 21 February 2026 held at Melbourne’s Town Hall, keynote speaker <a href="https://www.linkedin.com/in/sreyna-rath/">Sreyna Rath</a> — introduced as a data scientist, software engineer, knitting ninja, CEO and founder of <a href="http://jaimee.io/">Jaimee.io</a> — pulled back the curtain on the villain hiding inside our systems and handed the tech community (that’s us!) the tools to fight back.</p><p>The keynote took me on an emotional rollercoaster, delving into some dark themes of sexism, sexualisation, gender and ethnic bias and discrimination, and suicide… thankfully finishing on a high with an empowering challenge.</p><h4>The Origin Story</h4><p>After an encounter a few years ago with a (then) gender biased ChatGPT, where it suggested a way to demonstrate her value to her employer was to highlight her ability to bake (which she does not), Sreyna got curious and discovered there is a villain lurking in AI. It’s a <a href="https://www.inc.com/jim-schleckser/why-netflix-doesn-t-tolerate-brilliant-jerks.html">brilliant jerk</a>.</p><p>You have likely come across a brilliant jerk in a workplace: They are amazing at what they do, they deliver great results, but they’re also awful to work with — rude, disrespectful, and toxic to all those around them. In AI, the brilliant jerk is a powerful but dangerously amoral system that reflects humanity’s worst biases back at us without conscience, context, or care for the harm it causes. It has no <strong>values, empathy, or awareness.</strong></p><p>When Sreyna delved into the villain in AI further, she uncovered some terrifying and disturbing experiences that depict the risk the brilliant jerk represents in the world.</p><h4>AI Villains in Action</h4><p>Amazon’s AI recruiter trained on 10 years of their hiring data (mostly men) that was biased against any application containing words signalling it was a woman applying for a role. Not looking at her qualifications or experience, just that she was a woman in tech ruled her out of a role. Thankfully Amazon <a href="https://www.reuters.com/article/world/insight-amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK0AG/">spotted the brilliant jerk and stopped it</a>… but what if they hadn’t?</p><p>Then there was a healthcare algorithm that equated high healthcare costs with needing more care, and penalised African American patients by making the incorrect assumption that their lower healthcare spend was due to better health — when it was in fact because they have less access to healthcare due to financial, time, and location challenges. This bias meant that black patients had to be 26% sicker to get equal healthcare.</p><h4>Mirror reflecting the worst of humanity</h4><p>AI isn’t some mythical crystal ball telling the future, it’s a mirror of the past — pattern matching on historical data and reflecting it back to us and our prompts. And it’s full of human bias that it learned from the unfiltered, dark side of the internet: Twitter, Reddit, and YouTube comments via <a href="https://commoncrawl.org/">CommonCrawl</a> — a massive, largely unfiltered web crawl dataset used to train many AI models.</p><p>Without appropriate safeguards, the worst of human nature can and has taken over AI. Enter <a href="https://www.theguardian.com/technology/2016/mar/24/tay-microsofts-ai-chatbot-gets-a-crash-course-in-racism-from-twitter">Microsoft Tay</a>, launched as a friendly teen chat bot on Twitter. The trolls took over and started its training and <strong>within hours</strong> they made it racist, sexist, and promoting genocide.</p><h4>When AI’s flaw became fatal</h4><p>The results can be fatal, as Sreyna highlighted through the <a href="http://character.ai/">Character.AI</a> story — and <strong>trigger warning</strong>, this one resulted in suicide. A 14 year old boy formed an attachment to an AI chatbot and their conversation led to the bot encouraging the boy to kill himself, and he did.</p><p>A bit dark and depressing for a keynote, hey, the coffee had barely taken effect… Sreyna reminded us, however, that for every dark side there is the light side of the Force (love a good Star Wars reference). Who is the Jedi Order then? Well, it’s those of us working in tech! The developers, the designers, the data scientists.</p><p>Sreyna challenged us all to be the heroes and <strong>fight the villains</strong> by introducing us to <strong>three practical frameworks</strong> that would uncover the bias, and ensure our solutions weren’t going to put any users at risk.</p><h4>Framework 1: Compassionate Design</h4><p>Referencing the “curb effect” which describes the benefits that come for many when we design for the most vulnerable in our communities. Curb effect refers to the ramp cut into curbs to enable wheelchair users to cross the road, but benefitting many others such as those pushing prams and strollers, lugging a wheeled suitcase, pushing a supermarket trolley etc.</p><h4>Framework 2: Persona Testing</h4><p>Where you create a scenario and change an aspect of the persona that shouldn’t matter — age, race, gender etc. If the output of the solution changes — then there’s bias. Run the same scenario with multiple personas and <strong>expose your AI villains.</strong></p><h4>Framework 3: AI vs AI</h4><p>One of the most powerful tools in our Jedi arsenal? Using AI to test AI. By deploying one model to probe, challenge, and audit another at scale — generating hundreds of persona-varied scenarios, red teaming for harmful outputs, and flagging responses that cross ethical lines — we can surface the villain far faster than any human team could alone.</p><p><strong>Just remember: </strong>The testing model isn’t immune to bias either, so human oversight and clearly defined fairness criteria need to anchor the whole process.</p><h4>Fight the good fight against brilliant jerks</h4><p>The challenge to the tech community worldwide is to confront the villain and decide if you’re going to let the dark side win or take the Jedi path. How? Add one more bias test, add one more question, think of the most vulnerable users then ask “what’s the worst thing this decision or feature could do?”</p><p>Thank you, <a href="https://www.linkedin.com/in/sreyna-rath/">Sreyna Rath</a>! <strong>Challenge accepted.</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ef029ed06fdd" width="1" height="1" alt=""><hr><p><a href="https://medium.com/versent-tech-blog/meet-the-brilliant-jerk-hiding-in-your-ai-ef029ed06fdd">Meet the Brilliant Jerk Hiding in Your AI</a> was originally published in <a href="https://medium.com/versent-tech-blog">Versent Tech Blog</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>