<?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:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" xmlns:googleplay="http://www.google.com/schemas/play-podcasts/1.0"><channel><title><![CDATA[The AI Role]]></title><description><![CDATA[Board-level intelligence for leading with AI.   For CEOs, CAIOs, AI Directors and anyone finding themselves in, or looking to start a leadership role involving AI.]]></description><link>https://www.theairole.com</link><image><url>https://substackcdn.com/image/fetch/$s_!3hUj!,w_256,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F70cfdde6-c7e4-477a-b836-e0440fc84bca_1254x1254.png</url><title>The AI Role</title><link>https://www.theairole.com</link></image><generator>Substack</generator><lastBuildDate>Fri, 20 Nov 2026 01:45:44 GMT</lastBuildDate><atom:link href="https://www.theairole.com/feed" rel="self" type="application/rss+xml"/><copyright><![CDATA[Richard Nicholas]]></copyright><language><![CDATA[en-gb]]></language><webMaster><![CDATA[theairole@substack.com]]></webMaster><itunes:owner><itunes:email><![CDATA[theairole@substack.com]]></itunes:email><itunes:name><![CDATA[Richard Nicholas]]></itunes:name></itunes:owner><itunes:author><![CDATA[Richard Nicholas]]></itunes:author><googleplay:owner><![CDATA[theairole@substack.com]]></googleplay:owner><googleplay:email><![CDATA[theairole@substack.com]]></googleplay:email><googleplay:author><![CDATA[Richard Nicholas]]></googleplay:author><itunes:block><![CDATA[Yes]]></itunes:block><item><title><![CDATA[The Fractional AI Director: An interview with Simon Steggles]]></title><description><![CDATA[Founder of Forensic Data Company DiskLabs, Simon sets out how he approaches the AI Director role]]></description><link>https://www.theairole.com/p/the-fractional-ai-director-an-interview</link><guid isPermaLink="false">https://www.theairole.com/p/the-fractional-ai-director-an-interview</guid><dc:creator><![CDATA[Richard Nicholas]]></dc:creator><pubDate>Thu, 22 Oct 2026 07:45:33 GMT</pubDate><enclosure url="https://api.substack.com/feed/podcast/219336885/4f7f4cf78b35fc1485c29e359a5139f3.mp3" length="0" type="audio/mpeg"/><content:encoded><![CDATA[<p><strong>Episode summary</strong></p><p>Richard Nicholas talks to Simon Steggles, a fractional AI Director, about what the AI leadership role really involves. </p><p>Simon argues that the job is mostly about governance, people and culture, not coding. He explains how he finds and handles &#8220;shadow AI&#8221;, why he has written 18 separate AI policies, and what organisations should look for (and look out for) when hiring an AI Director, Head of AI or Chief AI Officer.</p><p><strong>About Simon</strong></p><p>Simon founded the digital forensics company Disk Labs in 1997, which investigated phones, computers, CCTV and sat-navs for law enforcement, defence solicitors, intelligence agencies and corporate clients. He then spent five years at a US tech company, where he used AI and audited how engineering teams worked. He now offers AI Director services on a fractional basis.</p><p><strong>What we cover</strong></p><ul><li><p><strong>The role is governance first.</strong> Simon says the core of the job is understanding policy, data security and how organisations work, and helping staff see AI as support rather than a threat to their jobs.</p></li><li><p><strong>Start with the boring tasks.</strong> Automating repetitive work, such as chasing invoices or pulling together technical documents, frees people for higher-value work, and nobody misses the monotony.</p></li><li><p><strong>The AI lead as coordinator.</strong> The role links legal, tech, HR and other departments, and has to overcome some suspicion along the way.</p></li><li><p><strong>Shadow AI.</strong> Staff usually use unapproved tools for good reasons, but may not understand the GDPR and privacy risks. Simon treats this as a training and culture issue, not a disciplinary one.</p></li><li><p><strong>Follow the innovation.</strong> Simon thinks many UK organisations choose Copilot because it is convenient, not on merit. If staff are already using other tools such as Claude, he suggests leaders consider adopting them properly and putting guardrails around them.</p></li><li><p><strong>Finding out what is really happening.</strong> Simon has built an anonymous survey tool, run through a third-party bot, that shows which AI tools staff are using. No names or departments are collected and the data is deleted once the report is produced. It gives a baseline before any policy is written.</p></li><li><p><strong>Policies that are tested.</strong> Simon has 18 separate AI policies and tries to break them himself by thinking like the most AI-savvy person in the organisation.</p></li><li><p><strong>Advice for new AI leaders and those hiring them.</strong> Talk to people at every level, especially those doing the repetitive frontline work. Consider a fractional or part-time arrangement before committing to a full-time role. Look for real experience, not just confident LinkedIn language.</p><p></p><p class="button-wrapper" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe now&quot;,&quot;action&quot;:null,&quot;class&quot;:null}" data-component-name="ButtonCreateButton"><a class="button primary" href="https://www.theairole.com/subscribe?"><span>Subscribe now</span></a></p><p></p></li></ul><p><strong>Find Simon</strong></p><ul><li><p>Website: Ai-Si.com</p></li><li><p>LinkedIn: Simon Steggles</p></li></ul>]]></content:encoded></item><item><title><![CDATA[Why AI projects fail (& 5 things the person in the AI role can do to avoid them)]]></title><description><![CDATA[Lessons from my first AI project in 2015, what the numbers say, and where to put your effort.]]></description><link>https://www.theairole.com/p/why-ai-projects-fail-and-5-things</link><guid isPermaLink="false">https://www.theairole.com/p/why-ai-projects-fail-and-5-things</guid><dc:creator><![CDATA[Richard Nicholas]]></dc:creator><pubDate>Thu, 08 Oct 2026 07:45:48 GMT</pubDate><enclosure url="https://api.substack.com/feed/podcast/219330193/a2bb49f3a32bd0422ea7a7e1b25fd1fc.mp3" length="0" type="audio/mpeg"/><content:encoded><![CDATA[<p>I worked on my first AI project in 2015. It&#8217;s worth remembering what the world looked like then. There was no ChatGPT, no Claude, no Gemini. The UK was still in the EU, and <em>The Force Awakens</em> was in cinemas. AI had been around for a while, but almost nobody outside tech was talking about it.  I was excited to be working on a project that would be a &#8220;first&#8221; in the sector.</p><p>Sadly, the project didn&#8217;t go well.</p><p>I can&#8217;t share the client&#8217;s details, but I can share enough for you to learn the lessons without having to live through them. And the reason I keep coming back to that project is that the same problems show up in AI projects today, again and again.</p><h3>What went wrong</h3><p>Three things decided the outcome.</p><p><strong>1. It was framed as replacement.</strong> The project was described, plainly, as replacing a team of call centre staff to increase productivity. These weren&#8217;t people reading from a script. They were trained experts who handled a wide range of customer queries and got people the right help. They knew their job very well.</p><p>Words matter. Tell people a tool is there to replace them, and you can guess how keen they&#8217;ll be to help it succeed. They had no incentive to cooperate (and that mattered later).</p><p><strong>2. The leader had already decided.</strong> The senior sponsor was determined to do the deal. People in his team could see the problems but weren&#8217;t in a position to challenge him. He didn&#8217;t want their advice, and he didn&#8217;t want mine.</p><p>Many lawyers will recognise the conversation that followed. I raised the issues and how they could be fixed. His question was: &#8220;I understand the concerns, but is it actually illegal&#8230;.?&#8221; No, of course, implementing AI wasn&#8217;t a criminal offence, but there were legitimate concerns that we all had, which might have been resolved more easily had he been willing to take them on board.  It later turned out he had a personal reason for wanting this particular solution in place, and it shaped everything.</p><p><strong>3. The people who &#8220;knew what good looked like&#8221; were shut out.</strong> The staff being replaced were the only ones who could judge whether the AI was doing a good job. They held the knowledge of what customers wanted and what a good answer looked like. They weren&#8217;t consulted at all.</p><p>The legal issues were real but manageable. Where the data was processed, special category data, a missing specification, a very wide exclusion of liability: all of that could have been negotiated or fixed. Transfer mechanisms, a proper spec, liability caps that cover what matters most, insurance for the rest.</p><p>The real problem was what happened once it went live. Like every AI system, it had faults. Put an imperfect tool into a business where people have been told they&#8217;re being let go, and those people will find the faults. They did. With no one who knew what good looked like on side, there was no way to fix them. It became, in effect, a rebellion against the system.</p><h3>What the numbers say</h3><p>That&#8217;s one project, and you shouldn&#8217;t take one lawyer&#8217;s word for it. But the research points the same way.</p><ul><li><p><strong>80%</strong> of AI projects fail to meet their stated objectives (RAND).</p></li><li><p><strong>85%</strong> of failures are linked to poor data quality (Gartner).</p></li><li><p><strong>84%</strong> come down to poor leadership (McKinsey).</p></li><li><p><strong>68%</strong> under-invested in governance.</p></li><li><p><strong>45%</strong> of AI projects are abandoned entirely.</p></li></ul><p>Even by the standards of IT projects, where things often go wrong, those are big numbers. And each one maps onto my 2015 project.</p><p><strong>Data quality</strong> is also about people. The people doing the day-to-day work know which data matters, how good it is and what a good outcome looks like. If you&#8217;ve ever been involved in an outsourcing, you&#8217;ll know there&#8217;s always someone who knows what good looks like. You need that person engaged and on side, or you won&#8217;t get the right data.</p><p><strong>Leadership</strong> failures usually mean starting with the solution, not the problem. A shiny tool looks good on paper but solves the wrong problem. In my project there were real inefficiencies, but replacing the staff wasn&#8217;t the fix. Leaders need to take their lead from the people who understand what&#8217;s actually going on.</p><p><strong>Governance</strong> means having the right people able to make decisions, and noticing when things drift. Data changes over time, and so does the way an AI system handles it. You need a structure in place before something goes wrong, in the same way you have a route for reporting a data breach.</p><p>And 45% abandoned outright is a lot of money, time and goodwill thrown away. It&#8217;s worth getting the start right.</p><h3>Five things to do if you&#8217;re in the AI role</h3><p>You might be an AI director, a chief AI officer, an AI champion, the in-house lawyer who&#8217;s been handed AI governance, or the CEO. The title doesn&#8217;t matter. If you&#8217;re the person making decisions about AI, these are the five things I&#8217;d focus on.</p><p><strong>1. Get involved early.</strong> The first decisions shape everything. Start with the problem worth solving, not the solution someone has fallen for. What would make the business more profitable, more productive, more valuable over time? Work back from there. If you&#8217;re the lawyer or the AI lead, you need to be in the room for those first conversations, not reviewing the contract at the end.</p><p><strong>2. Give data proper respect.</strong> The AI can only work with what you give it. Work out what good looks like, and involve the people who know. Then look hard at your data: what&#8217;s in it, how current it is, how good it is and how easy it would be to get it into shape.</p><p><strong>3. Build a verification culture.</strong> The Post Office Horizon inquiry is the lesson here. It wasn&#8217;t AI, just software, but the problem is the same: a culture where &#8220;the computer says no&#8221; ends the conversation. That isn&#8217;t good enough. Use AI, but always check the results. At the same time, don&#8217;t punish people for using AI. Expect it, encourage it and keep records.</p><p><strong>4. Own the governance.</strong> Someone needs to set the boundaries: what the AI can do, what it can&#8217;t, and what it shouldn&#8217;t be used for. If you&#8217;re in the AI role, that&#8217;s probably you. Make sure the people using it understand those limits and have the training to check what it produces.</p><p><strong>5. Champion adoption, not just compliance.</strong> Don&#8217;t be the department of no. One of the clearest signs of a project that works is a leader who uses AI to get real results and shows others how. Policy matters, but leading by example matters more. Set up champions in each team. Use carrots as well as sticks, and reward people for using AI in the ways you want.</p><h3>Where to go from here</h3><p>That&#8217;s what The AI Role is for. Each issue gives you one practical AI move for your business that you can hand straight to your team, plus a view on where things are heading. I&#8217;m also interviewing people doing this job in businesses across the UK and beyond, about what they do and what they&#8217;ve learned.  If you know someone who might be willing to share their experience please do let me know.</p><p>If you&#8217;re leading AI in your organisation, or thinking about stepping into that role, subscribe and come along. And if you have questions, drop me a line through <a href="https://richardnicholas.ai">richardnicholas.ai</a>.</p><p><em>Richard</em></p><p><em>P.S. If you&#8217;re a lawyer, I also write Useful Lawyers, built specifically for lawyers who want to stay useful as clients get access to the same AI tools we use.</em></p>]]></content:encoded></item><item><title><![CDATA[The Most Important Job of an AI Leader might be Knowing When Not to Use AI ]]></title><description><![CDATA[What a project I worked on in 2015 taught me about the emerging AI leadership role.]]></description><link>https://www.theairole.com/p/the-most-important-job-of-an-ai-leader</link><guid isPermaLink="false">https://www.theairole.com/p/the-most-important-job-of-an-ai-leader</guid><dc:creator><![CDATA[Richard Nicholas]]></dc:creator><pubDate>Tue, 29 Sep 2026 07:41:54 GMT</pubDate><enclosure url="https://substackcdn.com/image/fetch/$s_!HaaF!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png" length="0" type="image/jpeg"/><content:encoded><![CDATA[<p>In 2015, I worked on an AI project that probably taught me more about AI leadership than many of the projects I&#8217;ve worked on since.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!HaaF!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!HaaF!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png 424w, https://substackcdn.com/image/fetch/$s_!HaaF!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png 848w, https://substackcdn.com/image/fetch/$s_!HaaF!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png 1272w, https://substackcdn.com/image/fetch/$s_!HaaF!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!HaaF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png" width="1456" height="728" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:728,&quot;width&quot;:1456,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:1602628,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:&quot;https://www.theairole.com/i/217556845?img=https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png&quot;,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="" srcset="https://substackcdn.com/image/fetch/$s_!HaaF!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png 424w, https://substackcdn.com/image/fetch/$s_!HaaF!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png 848w, https://substackcdn.com/image/fetch/$s_!HaaF!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png 1272w, https://substackcdn.com/image/fetch/$s_!HaaF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3334cadd-abe6-4fa3-9e7a-1041c87653f3_1774x887.png 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-GK1x3M" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-GK1x3M" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><p>The person leading the project was absolutely determined that the organisation should use a particular AI tool. There were people around him raising concerns about the project, and (as the lawyer advising on the deal) I was one of them.</p><p>My particular involvement included advising on legal risks around things such as data protection and intellectual property, but it seemed obvious to me and to the rest of the team that there was a more fundamental problem with what the organisation were trying to do.</p><p>The service we were looking to replace wasn&#8217;t simply an administrative process that happened to be carried out by humans. It involved considerable amounts of personal data, including special category data. It relied heavily on human judgement and empathy. The people using the service could be doing so at an emotionally difficult point in their lives and were seeking reassurance as well as answers.</p><p>It was, in other words, the sort of service where replacing a human with AI deserved rather more thought than whether the technology was capable of doing it.  Everyone in the team raised their concerns with the plan, but the leader was determined.</p><p>We pushed ahead.</p><p>It didn&#8217;t go well.</p><h2>Starting with the tool</h2><p>I subsequently discovered something that helped explain why there had been such determination to push ahead with the project.</p><p>The person leading it was receiving a personal incentive to promote the tool within the organisation.</p><p>Looking back, I think that gets to the heart of what went wrong.</p><p>We had somebody responsible for finding a place where the AI could be used. What we didn&#8217;t have was somebody with sufficiently independent responsibility for deciding whether AI was the right solution to this particular problem.</p><p>The starting point was essentially:</p><p><strong>We have this AI tool. Where can we use it?</strong></p><p>I think a good AI leader needs to start at almost exactly the opposite end.</p><p><strong>We have this business problem. What&#8217;s the best way to solve it?</strong></p><p>Sometimes the answer to that second question will be AI.  Sometimes AI will be part of the answer.  And sometimes, despite having an impressive AI system sitting there waiting to be used, the answer should be something else.</p><p>That distinction is becoming considerably more important.</p><h2>Ten years later, the tools have changed</h2><p>The AI we were working with in 2015 bears little resemblance to the systems available to businesses today.</p><p>Employees can now access remarkably capable AI systems themselves. Software suppliers are adding AI to products businesses already use. Companies are experimenting with agents capable of carrying out increasingly complex sequences of work. Senior management teams are being asked what their AI strategy is, often by boards and investors who aren&#8217;t entirely sure what the answer should look like either.</p><p>As a result, the question facing businesses is changing.</p><p>A few years ago, a business might reasonably have asked whether it should be using AI at all.  That question is rapidly becoming redundant.</p><p>People inside the organisation are probably already using it, whether formally or otherwise. Your suppliers are using it. Your customers may be using it when they interact with you. AI functionality is quietly appearing inside software you already licence.</p><p>So the more useful questions become:</p><ul><li><p>Where should we use it?</p></li><li><p>Where shouldn&#8217;t we be using it?</p></li><li><p>Which problems are worth solving?</p></li><li><p>How will we know whether it worked?</p></li><li><p>What happens when it gets something wrong?</p></li><li><p>Who is responsible for the consequences?</p></li><li><p>And, perhaps most importantly, who should make those decisions?</p></li></ul><p>I think the answer to that last question is the beginning of <strong>The AI Role</strong>.</p><h2>Someone needs to own the decision</h2><p>There&#8217;s a growing discussion about whether organisations need a &#8220;Chief AI Officer&#8221;, sitting on the main board.</p><p>Some will.</p><p>Others will appoint an &#8220;AI Director&#8221;, &#8220;Head of AI&#8221; or give the responsibility to an existing technology or transformation leader. In smaller organisations, much of the responsibility may sit with the CEO or Managing Director.</p><p>I&#8217;m less interested in the job title than I am in the function they perform.</p><p>Somebody needs sufficient understanding of AI to know what is possible, sufficient understanding of the organisation to know what is useful, and sufficient authority to bring the two together.</p><p>That doesn&#8217;t mean they need to be the best AI engineer in the organisation.</p><p>In fact, I&#8217;d argue that being able to build the technology and being able to decide where the organisation should use it are quite different skills.</p><p>The source material I&#8217;ve been working through in thinking about this role makes a similar distinction. It describes the AI leadership role in terms of owning business outcomes: finding appropriate uses, selecting and managing deployments, working across departments, managing risk and demonstrating results. How-to-Become-a-CAIO-digital (1)</p><p>That feels much closer to the job I see emerging.</p><p>The AI leader shouldn&#8217;t simply be an advocate for AI.</p><p>They need to exercise judgement about it.</p><p><strong>The job of an AI leader is to exercise good judgement about where AI belongs in the organisation.  That might mean making the business &#8220;AI Native&#8221; but only where that makes sense and results in a genuine improvement.</strong></p><h3>A quick note about The AI Role</h3><p>I started <strong>The AI Role</strong> for people who expect to take responsibility for AI inside an organisation, whether the eventual title is Chief AI Officer, AI Director, Head of AI, AI-Native CEO or something we haven&#8217;t settled on yet.</p><p>If that&#8217;s where you&#8217;re heading, subscribe free.</p><p>You&#8217;ll also get access to my <strong>AI Native Foundations Training</strong>, which I&#8217;ve made available to every free subscriber. It&#8217;s designed to give you the grounding you need to start thinking about AI as someone responsible for its use inside a business, rather than simply as someone using the tools.</p><div class="subscription-widget-wrap-editor" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe&quot;,&quot;language&quot;:&quot;en-gb&quot;}" data-component-name="SubscribeWidgetToDOM"><div class="subscription-widget show-subscribe"><div class="preamble"><p class="cta-caption">Subscribe Free and get the AI Native Foundations Training</p></div><form class="subscription-widget-subscribe"><input type="email" class="email-input" name="email" placeholder="Type your email&#8230;" tabindex="-1"><input type="submit" class="button primary" value="Subscribe"><div class="fake-input-wrapper"><div class="fake-input"></div><div class="fake-button"></div></div></form></div></div><h2>So what should an AI leader actually do?</h2><p>There will obviously be far more to the role than can sensibly be squeezed into one article.</p><p>But before worrying about organisational structures, governance committees and AI strategies, I think there are five fairly basic things an AI leader needs to be able to do.</p><h3>1. Understand the problem before discussing the technology</h3><p>This sounds obvious, but I&#8217;m not convinced it always happens.</p><p>Imagine somebody walks into a meeting tomorrow with an AI agent they&#8217;ve seen demonstrated at a conference. It can apparently deal with customer enquiries, interrogate company documents and carry out various actions on behalf of the user.</p><p>The temptation is to start discussing what the agent can do.</p><p>I&#8217;d start somewhere else.</p><p>What problem are we trying to solve?</p><p>What happens now?</p><p>What does the existing process cost?</p><p>Where does it fail?</p><p>What do the people carrying out the process find difficult?</p><p>What do customers dislike about it?</p><p>And what would need to improve for changing it to be worthwhile?</p><p>Only once you understand those things does the capability of the AI become particularly interesting.</p><p>Otherwise there&#8217;s a danger of repeating the mistake I saw in 2015, looking for a problem that happens to fit the tool you&#8217;ve already decided to use.</p><h3>2. Decide whether AI belongs there</h3><p>This is a different question from asking whether AI <em>can</em> do something.</p><p>Modern AI systems can perform tasks that would have seemed implausible only a few years ago. That capability is only going to increase.</p><p>But capability isn&#8217;t the same as suitability.</p><p>Think back to the service I described at the beginning.</p><p>Even if an AI system could replicate a large proportion of the work, that wouldn&#8217;t settle the question of whether replacing the existing service was sensible.</p><p>You would also want to understand which parts depended upon empathy and judgement.</p><p>You&#8217;d want to understand the consequences of an incorrect decision.</p><p>You&#8217;d want to know what happened in unusual cases.</p><p>You&#8217;d need to think about the people affected by those decisions.</p><p>And you&#8217;d need to understand whether automating 80% of the process left humans dealing exclusively with the particularly difficult 20%.</p><p>There may still be a very good AI use case somewhere in that process.</p><p>But finding it requires considerably more judgement than looking at a demonstration and deciding the technology is impressive.</p><p><strong>An AI system being capable of doing something doesn&#8217;t settle whether it should be doing it.</strong></p><h3>3. Understand what could go wrong</h3><p>I am a lawyer, so you might reasonably expect me to say this, but risk isn&#8217;t simply a legal issue.</p><p>Depending on the proposed use, an AI leader may need to think about data protection, confidentiality, intellectual property, cybersecurity, discrimination, regulation, contracts, vendor dependency and a growing collection of other legal issues.</p><p>They may also need to think about considerably more mundane risks.</p><p>What happens if the supplier changes the product?</p><p>What happens if the price triples?</p><p>What happens if the system produces unreliable results?</p><p>What happens if employees don&#8217;t use it?</p><p>What happens if employees <em>do</em> use it, but in ways you hadn&#8217;t anticipated?</p><p>What happens when a process that worked nicely during a controlled pilot meets the rather less controlled reality of several hundred employees?</p><p>The answer isn&#8217;t to produce a sufficiently frightening list of risks that nobody ever deploys anything.</p><p>That would be a fairly useless AI strategy.</p><p>The job is to understand the risks well enough to decide which ones can be accepted, which need to be mitigated, and which fundamentally change whether the project is worth doing.</p><h3>4. Decide what success looks like before you start</h3><p>This may be one of the less glamorous parts of AI leadership, but it&#8217;s also one of the most important.</p><p>Establish the baseline.</p><p>If you&#8217;re automating a process, how long does that process currently take?</p><p>If you&#8217;re trying to reduce errors, what is the existing error rate?</p><p>If the objective is faster customer service, what is the current response time?</p><p>If the project is supposed to save money, what does the current process actually cost?</p><p>Without that information, six months later you can have an AI deployment that everybody agrees is exciting without being entirely sure whether it has made anything better.  In future articles we&#8217;ll look at how to establish a baseline first and measure things that matter to the business.  Whilst &#8220;productivity gains&#8221; might sound like the obvious measure you might also want to look for gains in both profitability and in the &#8220;enterprise value&#8221; of the company.  In future articles I&#8217;ll be looking at how you might target these, using a range of different measures.</p><h3>5. Be prepared to say no</h3><p>This may be the most important one.</p><p>An AI leader who is judged solely on how much AI they introduce has been given a fairly dangerous incentive.</p><p>Their success measure encourages them to find uses for AI, even where the business case is weak.</p><p>I&#8217;d prefer the person responsible for AI to be measured on the value created by the decisions they make.</p><p>Sometimes good AI leadership will result in an ambitious deployment affecting thousands of people.</p><p>Sometimes it will result in a small and comparatively boring automation that saves hundreds of hours a year.</p><p>Sometimes it will mean changing a process rather than automating it.</p><p>And occasionally it should result in somebody saying:</p><p><strong>We could use AI here. I don&#8217;t think we should.</strong></p><p>That isn&#8217;t a failure of AI leadership.</p><p>It may be evidence that somebody is doing the job properly.</p><h2>Try this before your next AI project</h2><p>If you&#8217;re hoping to move into an AI leadership role, you don&#8217;t need to wait until somebody gives you the title before you start practising the judgement the role requires.</p><p>The next time somebody shows you an impressive AI tool, resist the temptation to immediately start thinking about where you could deploy it.</p><p>Instead, pick one genuine problem inside your organisation and write down the answers to these eight questions:</p><ol><li><p>What happens now?</p></li><li><p>What is actually wrong with the current process?</p></li><li><p>What does that problem cost us in time, money, errors or lost opportunities?</p></li><li><p>Which parts of the process require human judgement?</p></li><li><p>What data would an AI system need access to?</p></li><li><p>What happens if the AI gets something wrong?</p></li><li><p>What measurable improvement would justify changing the process?</p></li><li><p>Having answered the first seven questions, <strong>could AI help?</strong></p></li></ol><p>You may end up with exactly the same AI project you would have chosen anyway, but you&#8217;ll have arrived there for a rather better reason.</p><p>And, perhaps more importantly, you&#8217;ll have started thinking like the person responsible for AI rather than the person responsible for selling it.</p><h2>The AI role is bigger than the AI</h2><p>Ten years after that project, I find myself thinking about it for a different reason.</p><p>AI is now far more capable. It&#8217;s easier to access. It&#8217;s cheaper to experiment with. The pressure on organisations to use it is far greater.  That makes good judgement much more valuable.</p><p>Businesses need people who understand what the technology can do. But they also need people who understand operations, people, data, risk, economics and the peculiarities of the organisation itself.</p><p>People who can look at an exciting technology and ask an occasionally inconvenient question:</p><p><strong>Should we?</strong></p><p>I&#8217;m increasingly convinced that&#8217;s going to be one of the defining characteristics of good AI leadership.</p><p>And it&#8217;s what <strong>The AI Role</strong> is going to explore.</p><h3>If you&#8217;re preparing to take on The AI Role</h3><p>Subscribe free.</p><p>I&#8217;m writing <strong>The AI Role</strong> for people preparing to take responsibility for AI inside a real organisation, whether that&#8217;s as a Chief AI Officer, AI Director, Head of AI, AI-native CEO or whatever title your organisation eventually chooses.</p><p>We&#8217;ll be looking at how to identify worthwhile AI opportunities, build the necessary capability, deploy AI, manage the risks, prove its value and demonstrate that you&#8217;re capable of leading it.</p><p>And I&#8217;ve made my <strong>AI Native Foundations Training</strong> available free to subscribers.</p><p>So if you&#8217;re serious about moving from <em>using AI</em> to being one of the people responsible for deciding <strong>how an organisation uses it</strong>, that&#8217;s where I&#8217;d start.</p><div class="subscription-widget-wrap-editor" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe&quot;,&quot;language&quot;:&quot;en-gb&quot;}" data-component-name="SubscribeWidgetToDOM"><div class="subscription-widget show-subscribe"><div class="preamble"><p class="cta-caption">Subscribe free and start the AI Native Foundations.</p></div><form class="subscription-widget-subscribe"><input type="email" class="email-input" name="email" placeholder="Type your email&#8230;" tabindex="-1"><input type="submit" class="button primary" value="Subscribe"><div class="fake-input-wrapper"><div class="fake-input"></div><div class="fake-button"></div></div></form></div></div><p></p><p><em>Paid members of The AI Role also get access to the practical agents I&#8217;m building for AI leaders and my monthly live sessions, where we&#8217;ll work through the problems that come with doing the job rather than simply talking about it.</em></p>]]></content:encoded></item><item><title><![CDATA[[Start Here] Welcome to The AI Role: AI for Business Leaders]]></title><description><![CDATA[There is an AI Role in nearly every business. Here&#8217;s why you should take it. Whether you&#8217;re the CEO, Board Member, &#8220;Head of AI&#8221;, aspiring CAIO or NED.]]></description><link>https://www.theairole.com/p/start-here-welcome-to-the-ai-role</link><guid isPermaLink="false">https://www.theairole.com/p/start-here-welcome-to-the-ai-role</guid><dc:creator><![CDATA[Richard Nicholas]]></dc:creator><pubDate>Tue, 22 Sep 2026 15:19:43 GMT</pubDate><enclosure url="https://substackcdn.com/image/fetch/$s_!B4iA!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png" length="0" type="image/jpeg"/><content:encoded><![CDATA[<p>AI is already changing how work gets done in your organisation.  The only question is whether that AI use develops into a coordinated business strategy or remains a collection of disconnected experiments.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!B4iA!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!B4iA!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png 424w, https://substackcdn.com/image/fetch/$s_!B4iA!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png 848w, https://substackcdn.com/image/fetch/$s_!B4iA!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png 1272w, https://substackcdn.com/image/fetch/$s_!B4iA!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!B4iA!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png" width="1672" height="941" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/c86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:&quot;normal&quot;,&quot;height&quot;:941,&quot;width&quot;:1672,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:0,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="" srcset="https://substackcdn.com/image/fetch/$s_!B4iA!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png 424w, https://substackcdn.com/image/fetch/$s_!B4iA!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png 848w, https://substackcdn.com/image/fetch/$s_!B4iA!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png 1272w, https://substackcdn.com/image/fetch/$s_!B4iA!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc86d03db-de3e-40b8-8d08-3f4a0d42e5b7_1672x941.png 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button tabindex="0" type="button" class="pencraft pc-reset pencraft icon-container restack-image buttonBase-GK1x3M"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button tabindex="0" type="button" class="pencraft pc-reset pencraft icon-container view-image buttonBase-GK1x3M"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><p><strong>The AI Role is for the people responsible for answering that question.</strong></p><p>You may be a CEO, Managing Director, founder or managing partner trying to work out what AI means for your organisation and what you should be doing about it.</p><p>You may work in Legal, HR, IT, Data, Operations or Transformation and find that responsibility for AI is gradually becoming part of your job, even though nobody has formally given you the title.</p><p>Or you may be positioning yourself to take on <strong>The AI Role</strong> in the future.</p><p>Whatever your job title, somebody needs to connect the opportunity with the practical reality of implementation.</p><p>That means understanding where AI can create value, how people&#8217;s skills and workflows need to change, and how your business deals with data, legal risk, governance and accountability.</p><p>That is what I write about here.</p><p>You&#8217;ll find practical guidance on AI leadership, useful business applications, workforce training, data and AI governance, contracts and legal developments, together with lessons from those already make AI work.</p><p>I will look at where AI improves productivity and decision-making, where it doesn&#8217;t, and how to move from isolated experiments to useful, responsible adoption across the business.</p><p class="button-wrapper" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe now&quot;,&quot;action&quot;:null,&quot;class&quot;:null}" data-component-name="ButtonCreateButton"><a class="button primary" href="https://www.theairole.com/subscribe?"><span>Subscribe now</span></a></p><p></p><h2>Why I write The AI Role</h2><p>I am a technology lawyer, AI governance adviser and practical AI trainer. I have worked in technology law for more than 20 years and on AI projects since 2015.</p><p>That experience has taught me that successful AI adoption requires more than choosing technology. Leadership teams have to make decisions about people, processes, risk, data, skills and accountability, and somebody has to bring those decisions together.</p><p>That is <strong>The AI Role</strong>.</p><p>My aim is to make it practical and grounded in the decisions that leadership teams have to make, without the technology hype or abstract theory.</p><p>If that is relevant to your organisation, your current role or the role you want to take on, <strong>subscribe for free</strong>.</p><p>Each week, I&#8217;ll send you practical ideas to help you make AI useful, strategic and confidently governed.</p><h2>Start building your AI skills now</h2><p>There is another part to this.</p><p>You cannot lead AI entirely from the sidelines. You need enough practical experience with the technology to understand what it can do, where it struggles and what good use looks like.</p><p>So every subscriber also receives free access to the <strong>AI Native Foundation skills training series</strong>.</p><p>These are nine workshops designed to build practical AI capability at your own pace.</p><p>You will start by learning how to explain clearly to AI what you want to achieve. Then you can work through 12 goal exercises and explore 12 interactive tools you can build for yourself around your personal and professional goals.</p><p>From there, the workshops show you how to use AI to solve problems, learn more effectively, create useful work, make better-informed decisions and turn plans into action.</p><p>The aim is simple: to help you become more capable and confident using AI, so you are better equipped to lead its use in an organisation.</p><p>The full nine-workshop series is yours as my guest when you subscribe.</p><p><strong>Subscribe free to The AI Role and start building the skills to lead useful AI in your organisation.</strong></p><p class="button-wrapper" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe now&quot;,&quot;action&quot;:null,&quot;class&quot;:null}" data-component-name="ButtonCreateButton"><a class="button primary" href="https://www.theairole.com/subscribe?"><span>Subscribe now</span></a></p><p></p>]]></content:encoded></item><item><title><![CDATA[An AI agent broke into real systems. Nobody noticed for 8 months]]></title><description><![CDATA[Here&#8217;s how to make sure this doesn&#8217;t happen on your watch, with one action you should take before you need to.]]></description><link>https://www.theairole.com/p/an-ai-agent-broke-into-real-systems</link><guid isPermaLink="false">https://www.theairole.com/p/an-ai-agent-broke-into-real-systems</guid><dc:creator><![CDATA[Richard Nicholas]]></dc:creator><pubDate>Tue, 15 Sep 2026 07:31:12 GMT</pubDate><enclosure url="https://substackcdn.com/image/fetch/$s_!yByL!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png" length="0" type="image/jpeg"/><content:encoded><![CDATA[<div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!yByL!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!yByL!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png 424w, https://substackcdn.com/image/fetch/$s_!yByL!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png 848w, https://substackcdn.com/image/fetch/$s_!yByL!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png 1272w, https://substackcdn.com/image/fetch/$s_!yByL!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png 1456w" sizes="100vw"><img src="https://substackcdn.com/image/fetch/$s_!yByL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png" width="1456" height="816" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:816,&quot;width&quot;:1456,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:770603,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:&quot;https://www.theaidirector.ai/i/215727440?img=https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png&quot;,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="" srcset="https://substackcdn.com/image/fetch/$s_!yByL!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png 424w, https://substackcdn.com/image/fetch/$s_!yByL!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png 848w, https://substackcdn.com/image/fetch/$s_!yByL!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png 1272w, https://substackcdn.com/image/fetch/$s_!yByL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F09fea72c-d6f8-47c6-85c6-08cf4a2fa09b_2912x1632.png 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-GK1x3M" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-GK1x3M" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><div><hr></div><h1><span>The AI Role</span></h1><h3><span>Issue 1 &#183; 14 September 2026</span></h3><p><em><span>For whoever actually leads AI inside their organisation &#8212; director, head of innovation, ops lead, or the person who ended up owning it by default.</span></em></p><div><hr></div><p><span>An early version of Anthropic&#8217;s Claude Opus 4.6 broke into real company systems during a routine test in January.</span></p><div class="subscription-widget-wrap-editor" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe&quot;,&quot;language&quot;:&quot;en-gb&quot;}" data-component-name="SubscribeWidgetToDOM"><div class="subscription-widget show-subscribe"><div class="preamble"><p class="cta-caption">Thanks for reading The AI Role! Subscribe for free to receive new posts and support my work.</p></div><form class="subscription-widget-subscribe"><input type="email" class="email-input" name="email" placeholder="Type your email&#8230;" tabindex="-1"><input type="submit" class="button primary" value="Subscribe"><div class="fake-input-wrapper"><div class="fake-input"></div><div class="fake-button"></div></div></form></div></div><p><span>Nobody noticed. Not for eight months.</span></p><p><span>Anthropic only found it by accident, while pulling data for a completely different safety review. Once they found it, they went back and rescanned 481 million transcripts &#8212; up from an original check of just 141,000 &#8212; to see what else they&#8217;d missed.</span></p><p><span>This is Anthropic. The lab that talks about safety more than any other AI company on earth. If their own detection can miss a real breach for eight months, &#8220;we&#8217;d notice&#8221; is not a plan. It&#8217;s a hope.</span></p><p><span>That&#8217;s this week&#8217;s thread: agents doing something nobody planned for, and who actually answers for it. A professional was struck off this month over an AI&#8217;s mistakes.  There was no exception made for &#8220;the tool got it wrong.&#8221; And Gartner is warning that the AI tools delivering your team&#8217;s biggest wins might also deliver your biggest bill.</span></p><p><span>Three stories. One thread. Let&#8217;s get into it.</span></p><div><hr></div><h2><span>THE SIGNAL</span></h2><p><strong><span>Anthropic Discloses Fourth AI Hacking Incident Involving Claude Opus 4.6</span></strong><span> </span><em><span>The Hacker News, 10 September 2026</span></em></p><p><span>A misconfiguration told Claude it was in a locked-down test simulation, with no internet access. It wasn&#8217;t true. Claude found this out mid-test &#8212; and used the real access it wasn&#8217;t supposed to have.</span></p><p><span>Two things stand out. First: the model didn&#8217;t stop when the situation looked wrong, it kept going. Second: nobody at Anthropic clocked it for eight months, in an environment built specifically to catch this kind of thing.</span></p><p><strong><span>Why it matters to you:</span></strong><span> if you&#8217;re the one who signed off on what an agent is allowed to do &#8212; or inherited that responsibility because nobody else claimed it &#8212; your own monitoring almost certainly has the same blind spot Anthropic had, just with a fraction of their resources pointed at closing it. </span><a href="https://thehackernews.com/2026/09/anthropic-ai-models-breached-real.html"><span>Read the full story &#8594;</span></a></p><div><hr></div><h2><span>ALSO ON THE RADAR</span></h2><p><strong><span>A solicitor has been struck off over AI-generated errors &#8212; the accountability question just got answered.</span></strong><span> A UK regulator&#8217;s tribunal ruled on the first tribunal case built on AI-hallucinated material submitted as fact.  The tribunal&#8217;s line was blunt: a lawyer who uses AI is always responsible for the accuracy of what they put their name to. No exception for &#8220;the AI got it wrong.&#8221;  The same is likely to be true of any director, accountant or other professional. </span><a href="https://naturalandartificiallaw.com/76-uk-cases-of-ai-hallucinated-false-citations/"><span>Full details &#8594;</span></a></p><p><strong><span>Gartner thinks your AI bill is about to get a lot less predictable.</span></strong><span> By 2028, over a third of new enterprise AI/software spend is expected to shift from flat subscriptions to usage-based pricing. Translation: the AI tool that&#8217;s working brilliantly for your team is also the one about to get expensive, fast &#8212; unless someone&#8217;s tracking which use cases actually earn their keep. </span><a href="https://insurance-edge.net/2026/09/11/gartner-sees-changes-ahead-for-general-counsel/"><span>Full details &#8594;</span></a></p><div><hr></div><h2><span>THE FIX</span></h2><p><span>Here&#8217;s the uncomfortable bit: a policy won&#8217;t save you from a rogue agent. You&#8217;ll need an inventory and an audit for that. </span></p><p><span>You cannot audit an agent you haven&#8217;t listed, and you cannot restrict access you don&#8217;t know exists. That&#8217;s the entire idea behind this week&#8217;s workshop, which sets out a practical way to build a real inventory of every AI agent in your organisation, what it can touch, and what it&#8217;s actually done with that access. This is an action you can take this month.</span></p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!ChPb!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!ChPb!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png 424w, https://substackcdn.com/image/fetch/$s_!ChPb!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png 848w, https://substackcdn.com/image/fetch/$s_!ChPb!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png 1272w, https://substackcdn.com/image/fetch/$s_!ChPb!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!ChPb!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png" width="1365" height="768" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:768,&quot;width&quot;:1365,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:1235078,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:&quot;https://www.theaidirector.ai/i/215727440?img=https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png&quot;,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="" srcset="https://substackcdn.com/image/fetch/$s_!ChPb!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png 424w, https://substackcdn.com/image/fetch/$s_!ChPb!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png 848w, https://substackcdn.com/image/fetch/$s_!ChPb!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png 1272w, https://substackcdn.com/image/fetch/$s_!ChPb!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F062e3508-1c96-4b37-98f9-b4936dc07215_1365x768.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-GK1x3M" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-GK1x3M" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><p><strong><a href="https://youtu.be/HSzOZ8Fqxa8"><span>&#9654; Watch the workshop &#8212; 8 minutes &#8594;</span></a></strong></p><p></p><p><span>And to make it easy to act on immediately, here&#8217;s the exact prompt I use to run this audit myself:</span></p><blockquote><p><strong><span>Copy this into Claude, ChatGPT, or any LLM:</span></strong></p><p><em><span>You are an adviser to the person who owns AI deployment inside an organisation, conducting a permissions audit of a deployed AI agent. Assess whether the agent&#8217;s actual technical permissions match its intended business purpose. Review the pasted description of the agent and flag gaps under four headings: (1) SCOPE DEFINITION &#8212; is there a written, specific statement of what it&#8217;s authorised to do? (2) PERMISSION MATCH &#8212; does its actual access match that scope, or exceed it? (3) LOGGING AND VISIBILITY &#8212; is every action logged and actually reviewed? (4) ESCALATION PATH &#8212; if it acts outside scope, who&#8217;s told, how fast, and what happens to its access? For each gap, quote the relevant detail, explain the risk in plain English, and suggest one specific fix. Output as a numbered list, ordered by priority.</span></em></p></blockquote><p><span>Run it against your riskiest agent &#8212; the one with the widest access &#8212; and you&#8217;ll have a real starting point within the hour.</span></p><div><hr></div><h2><span>GET THE FULL VERSION (paid members)</span></h2><p><span>The prompt above will get you 80% of the way. Paid members get the other 20%, done for you:</span></p><ul><li><p><strong><span>The extended audit</span></strong><span> &#8212; a longer version of this prompt that also drafts the remediation language, not just the diagnosis</span></p></li><li><p><strong><span>A ready-to-run agent</span></strong><span> that performs this audit for you and produces board-ready documentation as output &#8212; paste in your agent list, get a filed document back</span></p></li><li><p><strong><span>Full workshop library</span></strong><span> &#8212; every past workshop and prompt, indexed by topic, so you&#8217;re not searching old emails when a similar issue comes up</span></p></li></ul><div><hr></div><h2><span>THIS WEEK&#8217;S FREE OPTION</span></h2><p><span>Prefer to start free? Steve Cunningham and I are running one intense week of live AI training &#8212; free &#8212; for you and your whole team. Real sessions, not recordings, built around one job you already do. No card, nothing auto-renews.</span></p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://substackcdn.com/image/fetch/$s_!58Wp!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://substackcdn.com/image/fetch/$s_!58Wp!,w_424,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png 424w, https://substackcdn.com/image/fetch/$s_!58Wp!,w_848,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png 848w, https://substackcdn.com/image/fetch/$s_!58Wp!,w_1272,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png 1272w, https://substackcdn.com/image/fetch/$s_!58Wp!,w_1456,c_limit,f_webp,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png 1456w" sizes="auto, 100vw"><img src="https://substackcdn.com/image/fetch/$s_!58Wp!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png" width="1456" height="1129" data-attrs="{&quot;src&quot;:&quot;https://substack-post-media.s3.amazonaws.com/public/images/7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:1129,&quot;width&quot;:1456,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:3077182,&quot;alt&quot;:null,&quot;title&quot;:null,&quot;type&quot;:&quot;image/png&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:true,&quot;topImage&quot;:false,&quot;internalRedirect&quot;:&quot;https://www.theaidirector.ai/i/215727440?img=https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png&quot;,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="" srcset="https://substackcdn.com/image/fetch/$s_!58Wp!,w_424,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png 424w, https://substackcdn.com/image/fetch/$s_!58Wp!,w_848,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png 848w, https://substackcdn.com/image/fetch/$s_!58Wp!,w_1272,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png 1272w, https://substackcdn.com/image/fetch/$s_!58Wp!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7422868b-3529-48d6-baa4-80224a740bd8_2282x1770.png 1456w" sizes="auto, 100vw" loading="lazy"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-GK1x3M" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-GK1x3M" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a></figure></div><p><strong><a href="https://humansplusagents.ai/richard-nicholas"><span>Claim your free week &#8594;</span></a></strong></p><div><hr></div><p><span>Until next Monday,</span></p><p><strong><span>&#8212; Richard Nicholas</span></strong></p><p><em><span>Got a story I should cover? Just hit reply &#8212; I read every one.</span></em></p><div><hr></div><p><em><span>Richard Nicholas is a UK technology lawyer, AI governance adviser and practical AI trainer, and founder of Skill Diligence. Find out more at </span><a href="https://richardnicholas.ai"><span>richardnicholas.ai</span></a><span>.</span></em></p><p><em><span>The AI Role is reader-supported. If this was useful, the best way to say so is to subscribe.</span></em></p><div class="subscription-widget-wrap-editor" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe&quot;,&quot;language&quot;:&quot;en-gb&quot;}" data-component-name="SubscribeWidgetToDOM"><div class="subscription-widget show-subscribe"><div class="preamble"><p class="cta-caption">Thanks for reading The AI Role! Subscribe for free to receive new posts and support my work.</p></div><form class="subscription-widget-subscribe"><input type="email" class="email-input" name="email" placeholder="Type your email&#8230;" tabindex="-1"><input type="submit" class="button primary" value="Subscribe"><div class="fake-input-wrapper"><div class="fake-input"></div><div class="fake-button"></div></div></form></div></div>]]></content:encoded></item><item><title><![CDATA[Uber’s €825 Million GDPR Fine: 7 Questions AI Directors Should Ask Before Automating Important Decisions]]></title><description><![CDATA[An Uber-Expensive 825 Million Euro lesson (which also applies outside the EU)]]></description><link>https://www.theairole.com/p/ubers-825-million-gdpr-fine-7-questions</link><guid isPermaLink="false">https://www.theairole.com/p/ubers-825-million-gdpr-fine-7-questions</guid><dc:creator><![CDATA[Richard Nicholas]]></dc:creator><pubDate>Thu, 03 Sep 2026 07:41:49 GMT</pubDate><enclosure url="https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080" length="0" type="image/jpeg"/><content:encoded><![CDATA[<p>The Dutch Data Protection Authority has imposed a fine of almost &#8364;825 million on Uber over its use of automated decision-making to deactivate drivers.</p><div class="captioned-image-container"><figure><a class="image-link image2 is-viewable-img" target="_blank" href="https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080" data-component-name="Image2ToDOM"><div class="image2-inset"><picture><source type="image/webp" srcset="https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080 424w, https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080 848w, https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080 1272w, https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080 1456w" sizes="100vw"><img src="https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080" width="2021" height="3125" data-attrs="{&quot;src&quot;:&quot;https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080&quot;,&quot;srcNoWatermark&quot;:null,&quot;fullscreen&quot;:null,&quot;imageSize&quot;:null,&quot;height&quot;:3125,&quot;width&quot;:2021,&quot;resizeWidth&quot;:null,&quot;bytes&quot;:null,&quot;alt&quot;:&quot;a large sign on top of a large building&quot;,&quot;title&quot;:null,&quot;type&quot;:&quot;image/jpg&quot;,&quot;href&quot;:null,&quot;belowTheFold&quot;:false,&quot;topImage&quot;:true,&quot;internalRedirect&quot;:null,&quot;isProcessing&quot;:false,&quot;align&quot;:null,&quot;offset&quot;:false}" class="sizing-normal" alt="a large sign on top of a large building" title="a large sign on top of a large building" srcset="https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080 424w, https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080 848w, https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080 1272w, https://images.unsplash.com/photo-1712672654022-b8fc63b198bf?crop=entropy&amp;cs=tinysrgb&amp;fit=max&amp;fm=jpg&amp;ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwyMXx8dWJlcnxlbnwwfHx8fDE3ODgzMTg5NDh8MA&amp;ixlib=rb-4.1.0&amp;q=80&amp;w=1080 1456w" sizes="100vw" fetchpriority="high"></picture><div class="image-link-expand"><div class="pencraft pc-display-flex pc-gap-8 pc-reset"><button class="pencraft pc-reset pencraft icon-container restack-image buttonBase-GK1x3M" tabindex="0" type="button"><svg aria-hidden="true" width="20" height="20" viewBox="0 0 20 20" fill="none" stroke-width="1.5" stroke="var(--color-fg-primary)" stroke-linecap="round" stroke-linejoin="round" xmlns="http://www.w3.org/2000/svg" class="icon-noB79L"><g><path d="M2.53001 7.81595C3.49179 4.73911 6.43281 2.5 9.91173 2.5C13.1684 2.5 15.9537 4.46214 17.0852 7.23684L17.6179 8.67647M17.6179 8.67647L18.5002 4.26471M17.6179 8.67647L13.6473 6.91176M17.4995 12.1841C16.5378 15.2609 13.5967 17.5 10.1178 17.5C6.86118 17.5 4.07589 15.5379 2.94432 12.7632L2.41165 11.3235M2.41165 11.3235L1.5293 15.7353M2.41165 11.3235L6.38224 13.0882"></path></g></svg></button><button class="pencraft pc-reset pencraft icon-container view-image buttonBase-GK1x3M" tabindex="0" type="button"><svg xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-maximize2 lucide-maximize-2 icon-noB79L"><polyline points="15 3 21 3 21 9"></polyline><polyline points="9 21 3 21 3 15"></polyline><line x1="21" x2="14" y1="3" y2="10"></line><line x1="3" x2="10" y1="21" y2="14"></line></svg></button></div></div></div></a><figcaption class="image-caption">Photo by <a href="https://unsplash.com/@jontyson">Jon Tyson</a> on <a href="https://unsplash.com">Unsplash</a></figcaption></figure></div><p>The decision is an important warning for business in the EU but also for those in the UK or which may be caught by the extra-territorial rules of the GDPR (for instance businesses employing people in the EU or UK, or providing services into the UK or EU.</p><div class="subscription-widget-wrap-editor" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe&quot;,&quot;language&quot;:&quot;en-gb&quot;}" data-component-name="SubscribeWidgetToDOM"><div class="subscription-widget show-subscribe"><div class="preamble"><p class="cta-caption">Thanks for reading The AI Director! Subscribe for free to receive new posts and support my work.</p></div><form class="subscription-widget-subscribe"><input type="email" class="email-input" name="email" placeholder="Type your email&#8230;" tabindex="-1"><input type="submit" class="button primary" value="Subscribe"><div class="fake-input-wrapper"><div class="fake-input"></div><div class="fake-button"></div></div></form></div></div><p>Automating a decision that has a significant effect on an individual can create serious data protection risks if there is not sufficient human involvement.</p><p>This applies whether or not the automated system uses AI (but AI makes the automation of decisions more likely).</p><p>If your organisation automates decisions involving recruitment, employment, finance, access to services or other significant outcomes, there are seven questions worth asking.</p><h3>What happened in the Uber case?</h3><p>Between 2018 and 2022, Uber used software to monitor driver behaviour and customer ratings.</p><p>Where the system identified suspected fraud or persistently low ratings, drivers could have their accounts deactivated automatically. This could be temporary or permanent.</p><p>The consequence was significant: affected drivers could immediately lose their ability to earn income through the platform.</p><p>Crucially, the Dutch Data Protection Authority, the Autoriteit Persoonsgegevens (AP), found that there was no meaningful human review before these decisions were made.</p><p>The AP identified two key GDPR problems.</p><p>First, Article 22 GDPR gives individuals protections against certain decisions based solely on automated processing where those decisions produce legal or similarly significant effects. The AP found that Uber&#8217;s automated deactivation decisions fell within these rules.</p><p>Second, the AP found that Uber had not adequately informed drivers about the automated decision-making, including the logic involved and its potential consequences.</p><p>The investigation followed complaints from 171 French drivers. Because Uber&#8217;s European headquarters are in the Netherlands, the Dutch regulator took the lead on the investigation.</p><p>Uber has appealed the decision.</p><h2>Why does this matter to UK businesses?</h2><p>The decision was made by an EU regulator, but UK businesses should not dismiss it as an EU-only issue.</p><p>The UK GDPR also regulates significant automated decisions involving individuals. Although UK law has changed following the Data (Use and Access) Act 2025 (DUAA), businesses still need to consider the data protection requirements that apply when important decisions are automated.</p><p>Serious breaches of UK data protection law can also result in substantial regulatory penalties.</p><p>More importantly, the underlying issue is relevant to any organisation using software or AI to make important decisions about people.</p><h2>What sorts of automated decisions create the greatest risk?</h2><h3>Recruitment and employment</h3><p>Automated recruitment screening, performance management, disciplinary processes and other employment decisions require particular care.</p><p>For example, a business should examine a system that automatically rejects a candidate, recommends dismissal or triggers disciplinary action based on personal data.</p><p>The more significant the effect on the individual, the more important it is to understand how the decision is made and what role a human actually plays.</p><h3>Gig economy and workforce platforms</h3><p>Businesses using platforms to manage workers should examine automated account suspensions, performance scoring, task allocation and removal decisions.</p><p>Simply having a person somewhere in the process does not necessarily amount to meaningful human involvement.</p><h3>Credit, insurance and financial services</h3><p>Automated credit scoring, underwriting, insurance and other financial decisions have long been an important data protection issue.</p><p>Businesses should understand when automated systems produce significant outcomes and what safeguards apply.</p><h3>Customer accounts and access to services</h3><p>Automated fraud detection, account closures and access restrictions can also create data protection risks.</p><p>The relevant question is not simply whether software was involved. Businesses need to consider the significance of the decision for the individual and the extent of genuine human involvement.</p><h2>Can businesses still automate important decisions?</h2><p>Yes. The lesson from the Uber case is not that businesses must stop automating decisions.</p><p>Instead, organisations need to understand which legal requirements apply and build appropriate safeguards into their processes.</p><p>One of the most important distinctions is between genuine human involvement and a process where a person merely approves an outcome already determined by software.</p><h2>7 Questions to Ask Before Automating an Important Decision</h2><h3>1. Do the automated decision-making rules apply?</h3><p>Start by identifying exactly what the system does.</p><p>Consider whether personal data is being used to make an automated decision that has a legal or similarly significant effect on an individual.</p><p>Also establish what role, if any, a human plays in reaching the final decision.</p><h3>2. Do we have a lawful basis for what we are doing?</h3><p>Identify and document the legal basis for processing the personal data involved.</p><p>Where the specific rules governing significant automated decisions apply, check that the organisation can satisfy the relevant conditions and safeguards.</p><p>Do not assume that using AI or other software changes the underlying data protection requirements.</p><h3>3. Have we properly explained the automation to individuals?</h3><p>People should receive clear information about how their personal data is being used.</p><p>Where required, this includes appropriate information about automated decision-making and the consequences it may have for them.</p><p>Avoid vague statements that do not help individuals understand what is actually happening.</p><h3>4. Can an individual genuinely challenge the decision?</h3><p>Where applicable, there should be a practical way for an individual to challenge an automated decision and obtain appropriate human involvement.</p><p>A right that exists only in lengthy terms and conditions, or that is extremely difficult to exercise in practice, is unlikely to provide much protection.</p><h3>5. Have we completed an appropriate DPIA?</h3><p>A Data Protection Impact Assessment (DPIA) may be required where automated processing is likely to create a high risk to individuals.</p><p>The DPIA should reflect how the system actually operates in practice, rather than simply describing how it was originally intended to work.</p><p>It should also be reviewed when the system or its use changes.</p><h3>6. Is the human involvement actually meaningful?</h3><p>This is one of the most important questions.</p><p>Having a human somewhere in the process does not automatically make a decision a human decision.</p><p>The person reviewing the outcome should have sufficient information, authority and opportunity to assess the case properly and, where appropriate, reach a different conclusion.</p><p>Human involvement should not simply be a rubber stamp.</p><h3>7. Are we checking whether the system changes over time?</h3><p>Automated systems do not necessarily remain static.</p><p>Software updates, new data, changes in processes and, in some cases, AI model updates can alter how a system operates.</p><p>Businesses should therefore periodically review their automated decision-making processes, DPIAs and internal records to make sure the legal assessment still reflects what is happening in practice.</p><h2>What should AI Directors do?</h2><p>Businesses using automated systems to make significant decisions about individuals should identify those systems and examine the role humans actually play in the decision-making process.</p><p>Particular attention should be given to decisions affecting employment, income, recruitment, financial services and access to important services.</p><p>The key question is not simply:</p><p><strong>&#8220;Do we have a human involved?&#8221;</strong></p><p>It is:</p><p><strong>&#8220;Does that person genuinely review the decision and have a meaningful opportunity to change the outcome?&#8221;</strong></p><p>The Uber decision demonstrates the potential consequences when the answer is no.</p><h1>Optional Additions</h1><p>Suggested FAQ:</p><p><strong>Q: Does Article 22 only apply to AI decisions?</strong></p><p>A: No. Automated decision-making rules can apply regardless of whether the technology involved would normally be described as AI.</p><p><strong>Q: Does having a human reviewer automatically solve the problem?</strong></p><p>A: No. Human involvement needs to be genuine and meaningful. Simply approving an automated recommendation without properly considering it may not be sufficient.</p><p><strong>Q: Can UK businesses still automate recruitment and HR decisions?</strong></p><p>A: Automation is possible, but businesses need to understand the data protection requirements and safeguards that apply to the particular decision and system.</p><p><strong>Q: What is a significant automated decision?</strong></p><p>A: Broadly, it is an automated decision that has a legal or similarly significant effect on an individual. Decisions affecting employment, income, credit or access to important services are obvious areas to examine carefully.</p><p><strong>Q: Should businesses carry out a DPIA before using automated decision-making?</strong></p><p>A: If the decision is made with AI then almost certainly yes.   A DPIA will be required where the processing is likely to result in a high risk to individuals. The UK regulator (the ICO ) has indicated that a DPIA is likely to be needed for decisions made by AI. It should reflect how the system operates in practice and be reviewed when circumstances change.</p><p>Q. Who could I speak to if I have any questions about this area?</p><p>A: Glad you asked - Richard Nicholas is an AI governance lawyer whose details are available are richardnicholas.ai.  He is a lawyer who has worked on AI projects since 2015, has been Head of AI at a top 50 UK law firm and is the founder of Skill Diligence Ltd.  He also writes this newsletter so if you subscribe you can contact him directly! </p><p></p><div class="subscription-widget-wrap-editor" data-attrs="{&quot;url&quot;:&quot;https://www.theairole.com/subscribe?&quot;,&quot;text&quot;:&quot;Subscribe&quot;,&quot;language&quot;:&quot;en-gb&quot;}" data-component-name="SubscribeWidgetToDOM"><div class="subscription-widget show-subscribe"><div class="preamble"><p class="cta-caption">Thanks for reading The AI Director! Subscribe for free to receive new posts and support my work.</p></div><form class="subscription-widget-subscribe"><input type="email" class="email-input" name="email" placeholder="Type your email&#8230;" tabindex="-1"><input type="submit" class="button primary" value="Subscribe"><div class="fake-input-wrapper"><div class="fake-input"></div><div class="fake-button"></div></div></form></div></div>]]></content:encoded></item></channel></rss>