[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"node-by-slug-\u002Fen\u002Finsights\u002Fis-irrational-ai-making-our-decision-making-worse-en":3,"related-content-view-27069-en":415},["Reactive",4],{"entity":5,"errors":-1},{"__typename":6,"status":7,"metatag":8,"id":202,"path":203,"hideNavigation":204,"title":13,"translations":205,"sharedHeaderRefMedia":213,"sharedTeaserTitle":13,"sharedTeaserText":239,"sharedTeaserRefMedia":241,"sharedParagraphs":243,"sharedTaxrefIndustries":380,"sharedTaxrefExpertises":384,"sharedIntroText":388,"sharedCtaLink":204,"sharedDate":390,"sharedAuthor":392},"NodeInsight",true,[9,14,18,21,25,30,34,38,42,46,50,54,58,62,66,69,73,76,79,81,84,87,91,95,99,103,107,111,115,119,123,127,131,135,139,142,145,148,151,154,157,160,163,166,170,174,177,181,185,189,192,196,200],{"tag":10,"attributes":11},"meta",{"name":12,"content":13},"title","Is irrational AI making our decision-making worse?",{"tag":10,"attributes":15},{"name":16,"content":17},"description","In this Tech Tomorrow episode, David Elliman speaks with Stephanie Antonian about why AI hype can lead to decision paralysis and how leaders can bring AI back to measurable business value.",{"tag":10,"attributes":19},{"name":20,"content":17},"abstract",{"tag":10,"attributes":22},{"name":23,"content":24},"icbm","51.525772, -0.082716",{"tag":26,"attributes":27},"link",{"rel":28,"href":29},"canonical","https:\u002F\u002Fcms.zuehlke.com\u002Fen\u002Finsights\u002Fis-irrational-ai-making-our-decision-making-worse",{"tag":10,"attributes":31},{"name":32,"content":33},"robots","index, follow",{"tag":26,"attributes":35},{"rel":36,"href":37},"shortlink","https:\u002F\u002Fcms.zuehlke.com\u002Fapi\u002Fgraphql",{"tag":26,"attributes":39},{"rel":40,"href":41},"image_src","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Fmax_1300x1300\u002Fpublic\u002Fimages\u002FStephanie_Antonian_website_image.png?itok=jWMw-637,https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Fzuhlke_innovation_social.jpg",{"tag":10,"attributes":43},{"name":44,"content":45},"referrer","unsafe-url",{"tag":10,"attributes":47},{"name":48,"content":49},"rights","Copyright ©2026 | Zühlke Engineering AG | All rights reserved.",{"tag":10,"attributes":51},{"property":52,"content":53},"og:site_name","Zühlke Engineering AG",{"tag":10,"attributes":55},{"property":56,"content":57},"og:type","website",{"tag":10,"attributes":59},{"property":60,"content":61},"og:url","https:\u002F\u002Fwww.zuehlke.com\u002Fen\u002Finsights\u002Fis-irrational-ai-making-our-decision-making-worse",{"tag":10,"attributes":63},{"property":64,"content":65},"og:title","Is irrational AI making our decision-making worse? – Zühlke",{"tag":10,"attributes":67},{"property":68,"content":17},"og:description",{"tag":10,"attributes":70},{"property":71,"content":72},"og:image","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Fmax_1300x1300\u002Fpublic\u002Fimages\u002FStephanie_Antonian_website_image.png?itok=jWMw-637",{"tag":10,"attributes":74},{"property":71,"content":75},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Fzuhlke_innovation_social.jpg",{"tag":10,"attributes":77},{"property":78,"content":72},"og:image:url",{"tag":10,"attributes":80},{"property":78,"content":75},{"tag":10,"attributes":82},{"property":83,"content":72},"og:image:secure_url",{"tag":10,"attributes":85},{"property":83,"content":86},"https:\u002F\u002Fwww.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Fzuhlke_innovation_social.jpg",{"tag":10,"attributes":88},{"property":89,"content":90},"og:image:alt","image",{"tag":10,"attributes":92},{"property":93,"content":94},"og:updated_time","2026-06-15",{"tag":10,"attributes":96},{"property":97,"content":98},"place:location:longitude","7.438014",{"tag":10,"attributes":100},{"property":101,"content":102},"place:location:latitude","46.947944",{"tag":10,"attributes":104},{"property":105,"content":106},"og:street_address","80 Great Eastern Street",{"tag":10,"attributes":108},{"property":109,"content":110},"og:locality","London",{"tag":10,"attributes":112},{"property":113,"content":114},"og:region","England",{"tag":10,"attributes":116},{"property":117,"content":118},"og:postal_code","EC2A 3JL",{"tag":10,"attributes":120},{"property":121,"content":122},"og:country_name","United Kingdom",{"tag":10,"attributes":124},{"property":125,"content":126},"og:email","london@zuhlke.com",{"tag":10,"attributes":128},{"property":129,"content":130},"og:phone_number","+44 20 7113 5000",{"tag":10,"attributes":132},{"property":133,"content":134},"og:locale","en_CH",{"tag":10,"attributes":136},{"property":137,"content":138},"og:locale:alternate","de_CH",{"tag":10,"attributes":140},{"property":137,"content":141},"de_DE",{"tag":10,"attributes":143},{"property":137,"content":144},"en_DE",{"tag":10,"attributes":146},{"property":137,"content":147},"en_AT",{"tag":10,"attributes":149},{"property":137,"content":150},"de_AT",{"tag":10,"attributes":152},{"property":137,"content":153},"en_HK",{"tag":10,"attributes":155},{"property":137,"content":156},"en_SR",{"tag":10,"attributes":158},{"property":137,"content":159},"en_BG",{"tag":10,"attributes":161},{"property":137,"content":162},"en_SG",{"tag":10,"attributes":164},{"property":137,"content":165},"en_GB",{"tag":10,"attributes":167},{"property":168,"content":169},"article:publisher","https:\u002F\u002Fwww.facebook.com\u002Fzuehlke.group",{"tag":10,"attributes":171},{"property":172,"content":173},"article:published_time","2026-05-20",{"tag":10,"attributes":175},{"property":176,"content":94},"article:modified_time",{"tag":10,"attributes":178},{"name":179,"content":180},"twitter:card","summary_large_image",{"tag":10,"attributes":182},{"name":183,"content":184},"twitter:site","@zuehlke_group",{"tag":10,"attributes":186},{"name":187,"content":188},"twitter:site:id","247189965",{"tag":10,"attributes":190},{"name":191,"content":90},"twitter:image:alt",{"tag":10,"attributes":193},{"name":194,"content":195},"twitter:image","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Fmax_1300x1300\u002Fpublic\u002Fimages\u002FStephanie_Antonian_website_image.png?itok=jWMw-637,https:\u002F\u002Fwww.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Fzuhlke_innovation_social.jpg",{"tag":10,"attributes":197},{"name":198,"content":199},"google-site-verification","VEzolc1Vxy1KyRIlIqUSmVS7fqP_3_ZeBqPRby4g2mE",{"tag":201},"script","27069","\u002Fen\u002Finsights\u002Fis-irrational-ai-making-our-decision-making-worse",null,[206,209],{"path":203,"langcode":207},{"id":208},"en",{"path":210,"langcode":211},"\u002Fde\u002Finsights\u002Fverschlechtert-irrationale-ki-unsere-entscheidungsfindung",{"id":212},"de",{"id":214,"sharedMediaLegend":204,"mediaImage":215,"name":238},"173244",{"url":216,"width":217,"height":218,"alt":219,"title":204,"size":220,"mime":221,"variations":204,"headerMinimalVariations":222},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002FStephanie_Antonian_website_image.png",1280,720,"Tech Tomorrow podcast thumbnail featuring guest Stephanie Antonian smiling inside a circular frame on a purple and blue abstract background, with the text “Tech Tomorrow with David Elliman” displayed on the right.",1954169,"image\u002Fpng",[223,228,229,234,237],{"name":224,"url":225,"width":226,"height":227},"FOCAL_POINT32_SMALL","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_small\u002Fpublic\u002Fimages\u002FStephanie_Antonian_website_image.png.webp?itok=xVNioiuY",800,533,{"name":224,"url":225,"width":226,"height":227},{"name":230,"url":231,"width":232,"height":233},"FOCAL_POINT169_MEDIUM","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_16_9_medium\u002Fpublic\u002Fimages\u002FStephanie_Antonian_website_image.png.webp?itok=ba6FwKmC",1200,675,{"name":235,"url":236,"width":217,"height":218},"FOCAL_POINT169_LARGE","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_16_9_large\u002Fpublic\u002Fimages\u002FStephanie_Antonian_website_image.png.webp?itok=jGsrZJcK",{"name":235,"url":236,"width":217,"height":218},"Stephanie_Antonian_website_image.png",{"processed":240},"\u003Cp>\u003Cspan lang=\"EN-US\">In this Tech Tomorrow episode, David Elliman speaks with Stephanie Antonian about why AI hype can lead to decision paralysis and how leaders can bring AI back to measurable business value.\u003C\u002Fspan>\u003C\u002Fp>",{"id":214,"sharedMediaLegend":204,"mediaImage":242,"name":238},{"url":216,"width":217,"height":218,"alt":219,"title":204,"size":220,"mime":221,"variations":204},[244,251,258,264,270,297,302,348,353],{"__typename":245,"id":246,"pgEmbededCode":247,"fullWidth":250,"edgeToEdge":250},"ParagraphPgCustomCode","35114",{"value":248,"processed":248,"format":249},"\u003Ciframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless=\"\" src=\"https:\u002F\u002Fshare.transistor.fm\u002Fe\u002Fd15f4aca\">\u003C\u002Fiframe>","custom_code",false,{"__typename":252,"id":253,"ctaLink":204,"pgSharedText":254,"pgSharedTitle":204,"pgSharedCtaLink":204},"ParagraphPgText","35103",{"value":255,"processed":256,"format":257},"\u003Cp>AI is often sold as a route to sharper decisions, greater efficiency and more rational organisations. But what if the opposite is happening?\u003C\u002Fp>\u003Cp>In this episode of \u003Ca href=\"\u002Fnode\u002F26709\" target=\"_blank\" title=\"Discover all episodes\" data-entity-type=\"node\" data-entity-uuid=\"94f2836b-1d26-43d8-8630-29aaa98f52dc\" data-entity-substitution=\"canonical\">Tech Tomorrow\u003C\u002Fa>, David Elliman speaks with Stephanie Antonian, founder and CEO of Aestora, to explore a provocative question: is irrational AI making our decision-making worse?\u003C\u002Fp>\u003Cp>The conversation builds on Stephanie’s argument that much of what we call AI today is not reasoning in the classical sense, but probabilistic pattern-matching: useful in the right context, but risky when misunderstood.\u003C\u002Fp>","\u003Cp>AI is often sold as a route to sharper decisions, greater efficiency and more rational organisations. But what if the opposite is happening?\u003C\u002Fp>\u003Cp>In this episode of \u003Ca href=\"\u002Fen\u002Finsights\u002Ftech-tomorrow-podcast\" target=\"_blank\" title=\"Discover all episodes\" data-entity-type=\"node\" data-entity-uuid=\"94f2836b-1d26-43d8-8630-29aaa98f52dc\" data-entity-substitution=\"canonical\">Tech Tomorrow\u003C\u002Fa>, David Elliman speaks with Stephanie Antonian, founder and CEO of Aestora, to explore a provocative question: is irrational AI making our decision-making worse?\u003C\u002Fp>\u003Cp>The conversation builds on Stephanie’s argument that much of what we call AI today is not reasoning in the classical sense, but probabilistic pattern-matching: useful in the right context, but risky when misunderstood.\u003C\u002Fp>","full_html",{"__typename":252,"id":259,"ctaLink":204,"pgSharedText":260,"pgSharedTitle":263,"pgSharedCtaLink":204},"35104",{"value":261,"processed":262,"format":257},"\u003Cp>\u003Ca href=\"https:\u002F\u002Fwww.linkedin.com\u002Fin\u002Fstephanie-antonian-b5614843\u002F?originalSubdomain=uk\" target=\"_blank\" title=\"See Stephanie's profile\">Stephanie Antonian\u003C\u002Fa> is the founder and CEO of \u003Ca href=\"https:\u002F\u002Fwww.aestora.com\u002F\" target=\"_blank\" title=\"Visit Aestora's website\">Aestora\u003C\u002Fa>, the AI research lab behind the Digital Health Score. Her work explores the moral, philosophical and societal implications of artificial intelligence, with a particular focus on human dignity, decision-making and the stories we tell about technology.  \u003C\u002Fp>\u003Cp>Before founding Aestora, Stephanie worked with organisations including Google X, DeepMind, NASA, Harvard Innovation Lab and Accenture across business strategy, public policy, data science and product management. Her recent essay, ‘\u003Ca href=\"https:\u002F\u002Fwww.aestora.com\u002Fessays\u002Fai-is-irrational\" title=\"Read the essay\">AI is Irrational\u003C\u002Fa>’, argues that modern AI is often described in terms of intelligence while operating more like statistical guesswork than structured reasoning.\u003C\u002Fp>","\u003Cp>\u003Ca href=\"https:\u002F\u002Fwww.linkedin.com\u002Fin\u002Fstephanie-antonian-b5614843\u002F?originalSubdomain=uk\" target=\"_blank\" title=\"See Stephanie's profile\">Stephanie Antonian\u003C\u002Fa> is the founder and CEO of \u003Ca href=\"https:\u002F\u002Fwww.aestora.com\u002F\" target=\"_blank\" title=\"Visit Aestora's website\">Aestora\u003C\u002Fa>, the AI research lab behind the Digital Health Score. Her work explores the moral, philosophical and societal implications of artificial intelligence, with a particular focus on human dignity, decision-making and the stories we tell about technology.  \u003C\u002Fp>\u003Cp>Before founding Aestora, Stephanie worked with organisations including Google X, DeepMind, NASA, Harvard Innovation Lab and Accenture across business strategy, public policy, data science and product management. Her recent essay, ‘\u003Ca href=\"https:\u002F\u002Fwww.aestora.com\u002Fessays\u002Fai-is-irrational\" title=\"Read the essay\">AI is Irrational\u003C\u002Fa>’, argues that modern AI is often described in terms of intelligence while operating more like statistical guesswork than structured reasoning.\u003C\u002Fp>","Meet the guest: Stephanie Antonian",{"__typename":252,"id":265,"ctaLink":204,"pgSharedText":266,"pgSharedTitle":269,"pgSharedCtaLink":204},"35105",{"value":267,"processed":268,"format":257},"\u003Ch3>AI is not magic (and it is not always reasoning)\u003C\u002Fh3>\u003Cp>Stephanie’s central challenge is linguistic as much as technical. When organisations talk about AI as ‘intelligent’, they often imply that it reasons, understands and judges. But most modern AI systems are built on machine learning: they identify patterns in historical data and produce statistically likely outputs.\u003C\u002Fp>\u003Cp>That does not make them useless, but it does mean they are frequently misunderstood.\u003C\u002Fp>\u003Cp>For Stephanie, the distinction matters because business leaders may be expecting possibility from systems designed to produce probability. As she puts it in the episode, many organisations are ‘waiting for the final number of pi to be revealed’ — investing as if today’s systems will eventually become fully definable, logical, and certain.  \u003C\u002Fp>\u003Cp>In her view, leaders should stop asking, ‘What can AI do?’ and start asking, ‘Where does probability create value in our organisation?’\u003C\u002Fp>\u003Cp>That means identifying bounded use cases where pattern recognition is genuinely useful: forecasting demand, detecting anomalies, summarising large datasets, supporting software development, or improving operational workflows. It also means recognising where probabilistic outputs are not enough, especially in high-stakes decisions involving safety, fairness, accountability or human welfare.\u003C\u002Fp>\u003Ch3>The hype is creating decision paralysis\u003C\u002Fh3>\u003Cp>AI is meant to help organisations move faster. Yet Stephanie argues that in many cases it is doing the opposite. More information, more options, and more automated analysis can make it harder for leaders to act.\u003C\u002Fp>\u003Cp>Her diagnosis is blunt: AI is not necessarily making bad decisions for us; it may be making us avoid decisions altogether.  \u003C\u002Fp>\u003Cp>David connects this to a familiar engineering principle. Innovation usually benefits from focus, constraints, and a clear definition of what needs to be achieved. Too much abstraction can slow progress. Too many reports can obscure the product, service, or customer outcome that actually matters.\u003C\u002Fp>\u003Cp>This is where the AI conversation needs to become more grounded. When an organisation has launched pilots, created internal tools or experimented with generative AI, the key is to question whether those activities are translating into measurable business value.\u003C\u002Fp>\u003Cp>A \u003Ca href=\"https:\u002F\u002Fmlq.ai\u002Fmedia\u002Fquarterly_decks\u002Fv0.1_State_of_AI_in_Business_2025_Report.pdf\" target=\"_blank\" title=\"See the report\">2025 MIT Project NANDA report\u003C\u002Fa> found that despite significant enterprise investment in generative AI, only a small minority of initiatives were producing measurable value. This  finding has intensified debate around why so many AI pilots fail to scale.  \u003C\u002Fp>\u003Cp>We see the same issue in many enterprise AI programmes: \u003Ca href=\"\u002Fnode\u002F26993\" target=\"_blank\" title=\"Discover whyAI time savings aren’t translating into real AI business value\" data-entity-type=\"node\" data-entity-uuid=\"5016a102-4395-4db5-98cc-dba3eb037298\" data-entity-substitution=\"canonical\">time savings are useful, but they are not the same as business impact\u003C\u002Fa>. The real test is whether AI improves revenue, margin, risk reduction, quality, resilience, or customer outcomes.  \u003C\u002Fp>","\u003Ch3>AI is not magic (and it is not always reasoning)\u003C\u002Fh3>\u003Cp>Stephanie’s central challenge is linguistic as much as technical. When organisations talk about AI as ‘intelligent’, they often imply that it reasons, understands and judges. But most modern AI systems are built on machine learning: they identify patterns in historical data and produce statistically likely outputs.\u003C\u002Fp>\u003Cp>That does not make them useless, but it does mean they are frequently misunderstood.\u003C\u002Fp>\u003Cp>For Stephanie, the distinction matters because business leaders may be expecting possibility from systems designed to produce probability. As she puts it in the episode, many organisations are ‘waiting for the final number of pi to be revealed’ — investing as if today’s systems will eventually become fully definable, logical, and certain.  \u003C\u002Fp>\u003Cp>In her view, leaders should stop asking, ‘What can AI do?’ and start asking, ‘Where does probability create value in our organisation?’\u003C\u002Fp>\u003Cp>That means identifying bounded use cases where pattern recognition is genuinely useful: forecasting demand, detecting anomalies, summarising large datasets, supporting software development, or improving operational workflows. It also means recognising where probabilistic outputs are not enough, especially in high-stakes decisions involving safety, fairness, accountability or human welfare.\u003C\u002Fp>\u003Ch3>The hype is creating decision paralysis\u003C\u002Fh3>\u003Cp>AI is meant to help organisations move faster. Yet Stephanie argues that in many cases it is doing the opposite. More information, more options, and more automated analysis can make it harder for leaders to act.\u003C\u002Fp>\u003Cp>Her diagnosis is blunt: AI is not necessarily making bad decisions for us; it may be making us avoid decisions altogether.  \u003C\u002Fp>\u003Cp>David connects this to a familiar engineering principle. Innovation usually benefits from focus, constraints, and a clear definition of what needs to be achieved. Too much abstraction can slow progress. Too many reports can obscure the product, service, or customer outcome that actually matters.\u003C\u002Fp>\u003Cp>This is where the AI conversation needs to become more grounded. When an organisation has launched pilots, created internal tools or experimented with generative AI, the key is to question whether those activities are translating into measurable business value.\u003C\u002Fp>\u003Cp>A \u003Ca href=\"https:\u002F\u002Fmlq.ai\u002Fmedia\u002Fquarterly_decks\u002Fv0.1_State_of_AI_in_Business_2025_Report.pdf\" target=\"_blank\" title=\"See the report\">2025 MIT Project NANDA report\u003C\u002Fa> found that despite significant enterprise investment in generative AI, only a small minority of initiatives were producing measurable value. This  finding has intensified debate around why so many AI pilots fail to scale.  \u003C\u002Fp>\u003Cp>We see the same issue in many enterprise AI programmes: \u003Ca href=\"\u002Fen\u002Finsights\u002Fwhy-ai-time-savings-dont-create-business-value\" target=\"_blank\" title=\"Discover whyAI time savings aren’t translating into real AI business value\" data-entity-type=\"node\" data-entity-uuid=\"5016a102-4395-4db5-98cc-dba3eb037298\" data-entity-substitution=\"canonical\">time savings are useful, but they are not the same as business impact\u003C\u002Fa>. The real test is whether AI improves revenue, margin, risk reduction, quality, resilience, or customer outcomes.  \u003C\u002Fp>","Key takeaways from the episode",{"__typename":271,"id":272,"caption":204,"imageRight":250,"pgSharedTitle":273,"pgSharedCtaLink":204,"textMedia":274,"pgSharedText":295},"ParagraphPgTextMedia","35115","Human accountability cannot be automated away",{"id":275,"sharedMediaLegend":204,"mediaImage":276,"name":294},"173247",{"url":277,"width":278,"height":279,"alt":280,"title":204,"size":281,"mime":221,"variations":204,"textMediaVariations":282},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Firrational_AI.png",1048,1173,"Abstract 3D digital structure with glowing blue lines and purple geometric shapes, representing artificial intelligence, data networks and emerging technology.",2055924,[283,288,289,290,293],{"name":284,"url":285,"width":286,"height":287},"FIXED_WIDTH_SMALL","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffixed_width_small\u002Fpublic\u002Fimages\u002Firrational_AI.png.webp?itok=BWAP7JRh",900,1007,{"name":284,"url":285,"width":286,"height":287},{"name":284,"url":285,"width":286,"height":287},{"name":291,"url":292,"width":278,"height":279},"FIXED_WIDTH_MEDIUM","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffixed_width_medium\u002Fpublic\u002Fimages\u002Firrational_AI.png.webp?itok=pHvvcV0C",{"name":291,"url":292,"width":278,"height":279},"irrational_AI.png",{"value":296,"processed":296,"format":257},"\u003Cp>One of the strongest themes in the conversation is accountability. Stephanie argues that meaningful human control is not optional, because human beings remain responsible for the systems they create and deploy.\u003C\u002Fp>\u003Cp>That view aligns with the direction of AI regulation and governance. The \u003Ca href=\"https:\u002F\u002Fdigital-strategy.ec.europa.eu\u002Fen\u002Fpolicies\u002Fregulatory-framework-ai\" target=\"_blank\" title=\"Learn more about the EU AI Act\">EU AI Act\u003C\u002Fa> takes a risk-based approach, with stricter requirements for higher-risk systems, while frameworks such as the \u003Ca href=\"https:\u002F\u002Fwww.nist.gov\u002Fitl\u002Fai-risk-management-framework\" target=\"_blank\" title=\"Discover the NIST AI Risk Management Framework\">NIST AI Risk Management Framework\u003C\u002Fa> focus on governing, mapping, measuring and managing AI-related risks across the lifecycle.  \u003C\u002Fp>\u003Cp>But David makes an important distinction: organisations should not treat AI governance as an abstract ethics debate. They should treat it as safety engineering.\u003C\u002Fp>",{"__typename":252,"id":298,"ctaLink":204,"pgSharedText":299,"pgSharedTitle":204,"pgSharedCtaLink":204},"35116",{"value":300,"processed":301,"format":257},"\u003Cp>That means asking practical questions:\u003C\u002Fp>\u003Col>\u003Cli data-list-item-id=\"eaf7359bdf675997a62142b2b93423714\">Who is accountable for this AI-supported decision?  \u003C\u002Fli>\u003Cli data-list-item-id=\"e6483c6d10f4b766723ea18262a155f31\">Can we audit how the system reached its output?  \u003C\u002Fli>\u003Cli data-list-item-id=\"eb11bd298f46db327fb33d74efeeff2fd\">What must the system never do?  \u003C\u002Fli>\u003Cli data-list-item-id=\"e16b58f3238b6536ba9c1f9a9e94e2413\">Are we testing those constraints continuously?  \u003C\u002Fli>\u003Cli data-list-item-id=\"e9e9fcf0d10ebdadfb86ae07fec4fb88b\">Is there a human review point before the decision affects people, customers or operations?  \u003C\u002Fli>\u003C\u002Fol>\u003Cp>This is where responsible AI moves from policy to practice. Governance becomes part of the delivery system, not an afterthought. Our blogpost on \u003Ca href=\"\u002Fnode\u002F26279\" data-entity-type=\"node\" data-entity-uuid=\"3a589f8c-4c75-4707-a719-af1da804531a\" data-entity-substitution=\"canonical\" target=\"_blank\" title=\"Learn to mitigate AI risks with an AI governance framework\">AI governance frameworks\u003C\u002Fa> makes the same point: mature AI governance must cover the full lifecycle, from initial idea to deployment and ongoing monitoring.\u003C\u002Fp>\u003Ch3>AI value starts with cost, capability and context\u003C\u002Fh3>\u003Cp>When asked what business leaders should do first, Stephanie suggests a costing exercise.\u003C\u002Fp>\u003Cp>Before committing to AI at scale, leaders need to understand:\u003C\u002Fp>\u003Cul>\u003Cli data-list-item-id=\"e030d96b43ee9be40aca66200b00136a3\">What the technology costs today.\u003C\u002Fli>\u003Cli data-list-item-id=\"e7a41c04c3c96fa12510e580a1edc6d44\">What it is likely to cost in future.\u003C\u002Fli>\u003Cli data-list-item-id=\"e0b1663fa56beeb4b790493c6577ace94\">What its current capabilities really are.\u003C\u002Fli>\u003Cli data-list-item-id=\"ed2f8574056b82faa7c55c79af3ca0c03\">Where probabilistic systems create value.\u003C\u002Fli>\u003Cli data-list-item-id=\"e9967ab34dcdf1872efd437a694c405c5\">What benchmarks will prove whether experiments are worth scaling.  \u003C\u002Fli>\u003C\u002Ful>\u003Cp>This is a refreshingly practical position that shows that organisations do not need to become ‘AI-first’ to benefit from AI. In fact, they probably should not. They need to become outcome-first, \u003Ca href=\"\u002Fnode\u002F26994\" title=\"Discover how to choose AI initiatives that create real business value\" data-entity-type=\"node\" data-entity-uuid=\"1d488a1e-b0fb-4332-a9b1-6221ccebd74e\" data-entity-substitution=\"canonical\" target=\"_blank\">value-first\u003C\u002Fa>, and evidence-first.\u003C\u002Fp>\u003Cp>The foundation is not the model, but the \u003Ca href=\"\u002Fnode\u002F27000\" data-entity-type=\"node\" data-entity-uuid=\"71c4a910-7cd2-44f5-b2b7-e5afbede6efe\" data-entity-substitution=\"canonical\" target=\"_blank\" title=\"Discover what it really takes to scale AI from proof-of-concept to production\">operating context around the model\u003C\u002Fa>: data quality, integration, governance, workflow design, user adoption, and measurable ROI. Without those foundations, even impressive pilots can remain stuck in proof-of-concept mode.\u003C\u002Fp>\u003Ch3>The real risk is using AI to avoid hard questions\u003C\u002Fh3>\u003Cp>Perhaps the most uncomfortable point in the episode is that organisations may be using AI to avoid confronting deeper business problems.\u003C\u002Fp>\u003Cp>If customers are becoming poorer, AI will not automatically fix demand. If processes are overly complex, adding AI may make them even more complex. And if an organisation lacks strategic clarity, generative tools may simply produce more content, more analysis, and more noise.\u003C\u002Fp>\u003Cp>David captures this through the lens of engineering discipline: progress comes from knowing what needs to be built, why it matters, and how success will be tested. In software, quality is not created by wishing for it. It is built through architecture, tests, feedback loops and operational ownership.\u003C\u002Fp>\u003Cp>The same should apply to AI. The challenge is to resist both extremes: blind optimism and defensive inaction. AI is a powerful tool, but power without context can easily create complexity instead of value.\u003C\u002Fp>","\u003Cp>That means asking practical questions:\u003C\u002Fp>\u003Col>\u003Cli>Who is accountable for this AI-supported decision?  \u003C\u002Fli>\u003Cli>Can we audit how the system reached its output?  \u003C\u002Fli>\u003Cli>What must the system never do?  \u003C\u002Fli>\u003Cli>Are we testing those constraints continuously?  \u003C\u002Fli>\u003Cli>Is there a human review point before the decision affects people, customers or operations?  \u003C\u002Fli>\u003C\u002Fol>\u003Cp>This is where responsible AI moves from policy to practice. Governance becomes part of the delivery system, not an afterthought. Our blogpost on \u003Ca href=\"\u002Fen\u002Finsights\u002Fmitigate-ai-risks-with-an-ai-governance-framework\" data-entity-type=\"node\" data-entity-uuid=\"3a589f8c-4c75-4707-a719-af1da804531a\" data-entity-substitution=\"canonical\" target=\"_blank\" title=\"Learn to mitigate AI risks with an AI governance framework\">AI governance frameworks\u003C\u002Fa> makes the same point: mature AI governance must cover the full lifecycle, from initial idea to deployment and ongoing monitoring.\u003C\u002Fp>\u003Ch3>AI value starts with cost, capability and context\u003C\u002Fh3>\u003Cp>When asked what business leaders should do first, Stephanie suggests a costing exercise.\u003C\u002Fp>\u003Cp>Before committing to AI at scale, leaders need to understand:\u003C\u002Fp>\u003Cul>\u003Cli>What the technology costs today.\u003C\u002Fli>\u003Cli>What it is likely to cost in future.\u003C\u002Fli>\u003Cli>What its current capabilities really are.\u003C\u002Fli>\u003Cli>Where probabilistic systems create value.\u003C\u002Fli>\u003Cli>What benchmarks will prove whether experiments are worth scaling.  \u003C\u002Fli>\u003C\u002Ful>\u003Cp>This is a refreshingly practical position that shows that organisations do not need to become ‘AI-first’ to benefit from AI. In fact, they probably should not. They need to become outcome-first, \u003Ca href=\"\u002Fen\u002Finsights\u002Fhow-to-choose-ai-initiatives-that-create-real-business-value\" title=\"Discover how to choose AI initiatives that create real business value\" data-entity-type=\"node\" data-entity-uuid=\"1d488a1e-b0fb-4332-a9b1-6221ccebd74e\" data-entity-substitution=\"canonical\" target=\"_blank\">value-first\u003C\u002Fa>, and evidence-first.\u003C\u002Fp>\u003Cp>The foundation is not the model, but the \u003Ca href=\"\u002Fen\u002Finsights\u002Fscaling-ai-from-proof-of-concept-to-production\" data-entity-type=\"node\" data-entity-uuid=\"71c4a910-7cd2-44f5-b2b7-e5afbede6efe\" data-entity-substitution=\"canonical\" target=\"_blank\" title=\"Discover what it really takes to scale AI from proof-of-concept to production\">operating context around the model\u003C\u002Fa>: data quality, integration, governance, workflow design, user adoption, and measurable ROI. Without those foundations, even impressive pilots can remain stuck in proof-of-concept mode.\u003C\u002Fp>\u003Ch3>The real risk is using AI to avoid hard questions\u003C\u002Fh3>\u003Cp>Perhaps the most uncomfortable point in the episode is that organisations may be using AI to avoid confronting deeper business problems.\u003C\u002Fp>\u003Cp>If customers are becoming poorer, AI will not automatically fix demand. If processes are overly complex, adding AI may make them even more complex. And if an organisation lacks strategic clarity, generative tools may simply produce more content, more analysis, and more noise.\u003C\u002Fp>\u003Cp>David captures this through the lens of engineering discipline: progress comes from knowing what needs to be built, why it matters, and how success will be tested. In software, quality is not created by wishing for it. It is built through architecture, tests, feedback loops and operational ownership.\u003C\u002Fp>\u003Cp>The same should apply to AI. The challenge is to resist both extremes: blind optimism and defensive inaction. AI is a powerful tool, but power without context can easily create complexity instead of value.\u003C\u002Fp>",{"__typename":303,"id":304,"columns":305,"listItems":306,"text":345,"title":346,"variant":347},"ParagraphPgContentList","35111","3col",[307,317,324,331,338],{"id":308,"icon":309,"link":204,"logo":204,"media":204,"text":313,"title":316},"35106",{"id":310,"file":311},"170294",{"url":312},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002F2024-07\u002Fdouble-arrow-right-icon.svg",{"value":314,"processed":314,"format":315},"\u003Cp>Identify use cases where pattern recognition, prediction or summarisation genuinely improves an existing workflow.\u003C\u002Fp>","content_list","1. Define where probability is useful",{"id":318,"icon":319,"link":204,"logo":204,"media":204,"text":321,"title":323},"35107",{"id":310,"file":320},{"url":312},{"value":322,"processed":322,"format":315},"\u003Cp>Decide what success means in financial, operational, customer or risk terms before running the experiment.\u003C\u002Fp>","2. Set benchmarks before pilots begin",{"id":325,"icon":326,"link":204,"logo":204,"media":204,"text":328,"title":330},"35108",{"id":310,"file":327},{"url":312},{"value":329,"processed":329,"format":315},"\u003Cp>Build accountability, auditability, and human control into AI systems from the start.\u003C\u002Fp>","3. Treat governance as delivery infrastructure",{"id":332,"icon":333,"link":204,"logo":204,"media":204,"text":335,"title":337},"35109",{"id":310,"file":334},{"url":312},{"value":336,"processed":336,"format":315},"\u003Cp>Track whether AI improves outcomes, not just whether people use it or save time.\u003C\u002Fp>","4. Measure business value instead of activity",{"id":339,"icon":340,"link":204,"logo":204,"media":204,"text":342,"title":344},"35110",{"id":310,"file":341},{"url":312},{"value":343,"processed":343,"format":315},"\u003Cp>AI can support judgement, but it should not replace leadership, accountability, or strategic intent.\u003C\u002Fp>","5. Keep humans at the centre of creation","Stephanie’s advice is anti-delusion: leaders should experiment with AI, but with sharper questions and stronger evidence. A practical starting point would be:","What leaders should do next","icon",{"__typename":252,"id":349,"ctaLink":204,"pgSharedText":350,"pgSharedTitle":352,"pgSharedCtaLink":204},"35112",{"value":351,"processed":351,"format":257},"\u003Cp>The episode’s central question — is irrational AI making our decision-making worse? — does not have a simple yes or no answer.\u003C\u002Fp>\u003Cp>AI can improve decisions when it is used in the right context, with clear boundaries, measurable outcomes, and human accountability. But it can also create confusion when leaders expect certainty from probabilistic systems, strategy from pattern recognition, or transformation from tools that have not been integrated into real business workflows.\u003C\u002Fp>\u003Cp>The opportunity now is to make the AI conversation more rational: less hype, more evidence; fewer abstract promises, more practical engineering; less fascination with what AI might become, and more discipline around what it can deliver today.\u003C\u002Fp>\u003Cp>As organisations move from experimentation to implementation, that discipline may become the real competitive advantage.\u003C\u002Fp>","Evolving into a more rational AI conversation",{"__typename":354,"id":355,"title":356,"text":357,"pgCtaLink":358,"imageLayout":361,"media":362},"ParagraphPgTeaserBig","35113","Discover all Tech Tomorrow episodes","Tech Tomorrow is your front-row seat to the conversations redefining the future. Each episode explores one big question about data, AI, or emerging tech, giving leaders clear, focused answers they can trust.",{"title":359,"url":360,"internal":7},"Explore all episodes","\u002Fen\u002Finsights\u002Ftech-tomorrow-podcast","left",{"id":363,"sharedMediaLegend":204,"mediaImage":364,"name":379},"173199",{"url":365,"width":366,"height":366,"alt":367,"title":204,"size":368,"mime":221,"variations":204,"teaserVariations":369},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Ftech_tomorrow_teaser_blogs.png",1080,"Tech Tomorrow podcast thumbnail featuring host David Elliman smiling inside a circular frame on a purple and blue abstract background, with the text “Tech Tomorrow with David Elliman” displayed on the right.",3077352,[370,372,373,377,378],{"name":224,"url":371,"width":226,"height":227},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_small\u002Fpublic\u002Fimages\u002Ftech_tomorrow_teaser_blogs.png.webp?itok=_fFx-z4d",{"name":224,"url":371,"width":226,"height":227},{"name":374,"url":375,"width":366,"height":376},"FOCAL_POINT54_MEDIUM","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_5_4_medium\u002Fpublic\u002Fimages\u002Ftech_tomorrow_teaser_blogs.png.webp?itok=IjFixnwC",864,{"name":374,"url":375,"width":366,"height":376},{"name":374,"url":375,"width":366,"height":376},"tech_tomorrow_teaser_blogs.png",[381],{"id":382,"name":383},"855","All industries",[385],{"id":386,"name":387},"110","AI",{"processed":389},"\u003Cp>In this episode of the Tech Tomorrow podcast, David Elliman speaks with Stephanie Antonian, founder and CEO of Aestora, about the limits of machine learning, the danger of AI hype, and why leaders need clearer benchmarks before they scale AI.\u003C\u002Fp>\n",{"timestamp":391},1781524800,[393],{"status":7,"contactName":394,"contactProfession":395,"contactLinkedinUrl":396,"sharedEmail":399,"sharedRefMedia":400},"David Elliman","Global Chief of Software Engineering",{"title":397,"url":398,"internal":250},"","https:\u002F\u002Fwww.linkedin.com\u002Fin\u002Felliman\u002F","david.elliman@zuhlke.com",{"id":401,"sharedMediaLegend":204,"mediaImage":402,"name":414},"167342",{"url":403,"width":226,"height":226,"alt":394,"title":204,"size":404,"mime":405,"variations":204,"authorVariations":406},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Fdavid-elliman_0.jpeg",100830,"image\u002Fjpeg",[407,410,411,412,413],{"name":408,"url":409,"width":226,"height":226},"FOCAL_POINT11_SMALL","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_1_1_small\u002Fpublic\u002Fimages\u002Fdavid-elliman_0.jpeg.webp?itok=KqGd6fb7",{"name":408,"url":409,"width":226,"height":226},{"name":408,"url":409,"width":226,"height":226},{"name":408,"url":409,"width":226,"height":226},{"name":408,"url":409,"width":226,"height":226},"david-elliman.jpeg",[416,449,472,500],{"id":417,"path":418,"sharedTeaserTitle":419,"sharedTeaserRefMedia":420,"sharedTaxrefIndustries":439,"sharedTaxrefExpertises":441},"26577","\u002Fen\u002Finsights\u002Fzenai-solving-the-challenge-of-secure-scalable-enterprise-ai","ZenAI: Solving the challenge of secure, scalable enterprise AI",{"id":421,"sharedMediaLegend":204,"mediaImage":422,"name":438},"171958",{"url":423,"width":424,"height":425,"alt":426,"title":204,"size":427,"mime":405,"variations":204,"relatedContentVariations":428},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Fblog_ZenAI.jpg",1024,768,"Stylised chatbot icon glowing with digital data, representing ZenAI as an enterprise AI solution.",270941,[429,431,432,436,437],{"name":224,"url":430,"width":226,"height":227},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_small\u002Fpublic\u002Fimages\u002Fblog_ZenAI.jpg.webp?itok=-wJunfdU",{"name":224,"url":430,"width":226,"height":227},{"name":433,"url":434,"width":424,"height":435},"FOCAL_POINT32_MEDIUM","https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_medium\u002Fpublic\u002Fimages\u002Fblog_ZenAI.jpg.webp?itok=dCD9EB5A",683,{"name":433,"url":434,"width":424,"height":435},{"name":433,"url":434,"width":424,"height":435},"blog_ZenAI.jpg",[440],{"id":382,"name":383},[442,443,446],{"id":386,"name":387},{"id":444,"name":445},"1048","Cybernetic Delivery",{"id":447,"name":448},"1063","ZenAI",{"id":450,"path":451,"sharedTeaserTitle":452,"sharedTeaserRefMedia":453,"sharedTaxrefIndustries":468,"sharedTaxrefExpertises":470},"7862","\u002Fen\u002Finsights\u002Fresponsible-ai-framework-for-ethical-applications","Responsible AI: a framework for ethical solutions ",{"id":454,"sharedMediaLegend":204,"mediaImage":455,"name":467},"140398",{"url":456,"width":217,"height":218,"alt":457,"title":204,"size":458,"mime":405,"variations":204,"relatedContentVariations":459},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Fengineer-examining-robotic-arm-in-office.jpg","Female engineer examining a robotic arm in an office",154453,[460,462,463,465,466],{"name":224,"url":461,"width":226,"height":227},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_small\u002Fpublic\u002Fimages\u002Fengineer-examining-robotic-arm-in-office.jpg.webp?itok=RRMCRWpT",{"name":224,"url":461,"width":226,"height":227},{"name":433,"url":464,"width":366,"height":218},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_medium\u002Fpublic\u002Fimages\u002Fengineer-examining-robotic-arm-in-office.jpg.webp?itok=UeWo-A-E",{"name":433,"url":464,"width":366,"height":218},{"name":433,"url":464,"width":366,"height":218},"engineer-examining-robotic-arm-in-office.jpg",[469],{"id":382,"name":383},[471],{"id":386,"name":387},{"id":473,"path":474,"sharedTeaserTitle":475,"sharedTeaserRefMedia":476,"sharedTaxrefIndustries":493,"sharedTaxrefExpertises":495},"25818","\u002Fen\u002Finsights\u002Ffrom-code-to-conversation-the-new-era-of-digital-interaction","Digital interaction: the role of LLMs in personalising experiences",{"id":477,"sharedMediaLegend":204,"mediaImage":478,"name":492},"169796",{"url":479,"width":480,"height":481,"alt":482,"title":204,"size":483,"mime":405,"variations":204,"relatedContentVariations":484},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Fai-touch-screen_1.jpg",1920,1617,"\"\"",333374,[485,487,488,490,491],{"name":224,"url":486,"width":226,"height":227},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_small\u002Fpublic\u002Fimages\u002Fai-touch-screen_1.jpg.webp?itok=ftqe8Ug0",{"name":224,"url":486,"width":226,"height":227},{"name":433,"url":489,"width":232,"height":226},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_medium\u002Fpublic\u002Fimages\u002Fai-touch-screen_1.jpg.webp?itok=UGyvSLba",{"name":433,"url":489,"width":232,"height":226},{"name":433,"url":489,"width":232,"height":226},"ai-touch-screen.jpg",[494],{"id":382,"name":383},[496,497],{"id":386,"name":387},{"id":498,"name":499},"129","Experience design",{"id":501,"path":502,"sharedTeaserTitle":503,"sharedTeaserRefMedia":504,"sharedTaxrefIndustries":520,"sharedTaxrefExpertises":524},"25745","\u002Fen\u002Finsights\u002Fai-in-software-development-lessons-for-business-leaders","AI software development: what business leaders should know",{"id":505,"sharedMediaLegend":204,"mediaImage":506,"name":519},"5868",{"url":507,"width":217,"height":508,"alt":509,"title":204,"size":510,"mime":405,"variations":204,"relatedContentVariations":511},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fimages\u002Ffront-end_engineer_blog.jpg",854,"Woman front-end engineer at work at Zühlke ",120084,[512,514,515,517,518],{"name":224,"url":513,"width":226,"height":227},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_small\u002Fpublic\u002Fimages\u002Ffront-end_engineer_blog.jpg.webp?itok=TRqY4Ld3",{"name":224,"url":513,"width":226,"height":227},{"name":433,"url":516,"width":232,"height":226},"https:\u002F\u002Fcms.zuehlke.com\u002Fsites\u002Fdefault\u002Ffiles\u002Fstyles\u002Ffocal_point_3_2_medium\u002Fpublic\u002Fimages\u002Ffront-end_engineer_blog.jpg.webp?itok=DjHZPbTs",{"name":433,"url":516,"width":232,"height":226},{"name":433,"url":516,"width":232,"height":226},"front-end_engineer_blog.jpg",[521],{"id":522,"name":523},"854","Other industries",[525,526,529],{"id":386,"name":387},{"id":527,"name":528},"891","New technologies",{"id":530,"name":531},"21","Software engineering"]