https://thesuperskills.com/research Research on what increasingly capable AI does to human capability, judgement, learning and work. Updated as pages are added and reviewed. en-GB Wed, 26 Aug 2026 23:53:25 GMT https://thesuperskills.com/research/usage-theatrehttps://thesuperskills.com/research/usage-theatreWed, 26 Aug 2026 09:00:00 GMTUsage Theatre: what organisations perform when they measure AI use instead of AI value. Defined by Rahim Hirji. https://thesuperskills.com/research/the-verifiers-discounthttps://thesuperskills.com/research/the-verifiers-discountWed, 26 Aug 2026 09:00:00 GMTThe Verifier's Discount: what happens to the value of human work when the machine produces and the human checks. Defined by Rahim Hirji. https://thesuperskills.com/research/the-unclaimed-hourhttps://thesuperskills.com/research/the-unclaimed-hourWed, 26 Aug 2026 09:00:00 GMTThe Unclaimed Hour: the capacity AI creates that nobody decides how to use, so drift decides. Defined by Rahim Hirji. https://thesuperskills.com/research/neither-ai-hype-nor-doomhttps://thesuperskills.com/research/neither-ai-hype-nor-doomWed, 26 Aug 2026 09:00:00 GMTWhy both AI enthusiasm and AI pessimism lead to bad decisions, and how a third posture, grounded in realistic assessment and human capability, outperforms either. https://thesuperskills.com/research/the-superskills-thesishttps://thesuperskills.com/research/the-superskills-thesisWed, 26 Aug 2026 09:00:00 GMTWhy human capability is the only sustainable AI strategy. Tools commoditise; capability compounds. The foundational argument behind SuperSkills. https://thesuperskills.com/research/ai-glossaryhttps://thesuperskills.com/research/ai-glossaryWed, 26 Aug 2026 09:00:00 GMTThe vocabulary of drift versus design: capability debt, synthetic seniority, the missing rungs, the Verifier's Discount and more, defined by Rahim Hirji. https://thesuperskills.com/research/missing-rungshttps://thesuperskills.com/research/missing-rungsWed, 26 Aug 2026 09:00:00 GMTWhy automating junior tasks creates a leadership-pipeline risk most organisations miss until it is too late, and how to build new rungs on purpose. https://thesuperskills.com/research/the-missed-reps-of-ai-seniorityhttps://thesuperskills.com/research/the-missed-reps-of-ai-seniorityWed, 26 Aug 2026 09:00:00 GMTAI does not raise the bar for humans so much as remove the reps that made them good in the first place. Why the missed reps are the cost no one is measuring. https://thesuperskills.com/research/superskill-augmented-mindsethttps://thesuperskills.com/research/superskill-augmented-mindsetWed, 26 Aug 2026 09:00:00 GMTThe culminating SuperSkill: working with AI so it amplifies rather than atrophies human capability, what the evidence shows, and how the partnership succeeds or fails. https://thesuperskills.com/research/ai-reading-listhttps://thesuperskills.com/research/ai-reading-listWed, 26 Aug 2026 09:00:00 GMTA gap-mapped approach to AI reading for leadership teams. The Reading Compass, a four-quadrant shortlist, and how to convert reading into shared language and decisions. https://thesuperskills.com/research/ai-readiness-liehttps://thesuperskills.com/research/ai-readiness-lieWed, 26 Aug 2026 09:00:00 GMTWhy treating AI readiness as a technology problem builds on a foundation that cannot hold, and the five-link readiness chain, strategy, data, people, process, governance, that actually decides it. https://thesuperskills.com/research/synthetic-seniorityhttps://thesuperskills.com/research/synthetic-seniorityWed, 26 Aug 2026 09:00:00 GMTSynthetic seniority is when AI enables junior professionals to produce senior-looking output while the judgement, pattern recognition and contextual wisdom of real seniority has not been built. https://thesuperskills.com/research/superskill-principled-innovationhttps://thesuperskills.com/research/superskill-principled-innovationWed, 26 Aug 2026 09:00:00 GMTWhy innovation without ethical grounding destroys long-term value, what the evidence shows, and why the capacity grows more valuable as technology's power increases. https://thesuperskills.com/research/superskill-introductionhttps://thesuperskills.com/research/superskill-introductionWed, 26 Aug 2026 09:00:00 GMTThe seven human capabilities that decide who thrives as AI absorbs cognitive work, and why they compound while tools commoditise. https://thesuperskills.com/research/superskill-global-adaptabilityhttps://thesuperskills.com/research/superskill-global-adaptabilityWed, 26 Aug 2026 09:00:00 GMTWhy global adaptability is a learnable meta-capability, what the evidence shows, and how AI can both amplify and remove the struggle that builds it. https://thesuperskills.com/research/four-generations-and-aihttps://thesuperskills.com/research/four-generations-and-aiWed, 26 Aug 2026 09:00:00 GMTSkills as dignity, technology as a stress test, responsibility as inheritance. The SuperSkills Ladder, told through four generations of disruption. https://thesuperskills.com/research/superskill-empathyhttps://thesuperskills.com/research/superskill-empathyWed, 26 Aug 2026 09:00:00 GMTWhy empathy concentrates in value as AI can simulate it, how it works, where it fails, and why people still want it from humans. https://thesuperskills.com/research/design-versus-drifthttps://thesuperskills.com/research/design-versus-driftWed, 26 Aug 2026 09:00:00 GMTMost AI harm comes not from bad actors but from organisations that never made a choice. The drift-versus-design framework, human capability debt, and how to tell where you are. https://thesuperskills.com/research/decision-qualityhttps://thesuperskills.com/research/decision-qualityWed, 26 Aug 2026 09:00:00 GMTThe quiet failure mode of AI-assisted decisions: teams stop thinking and start approving. How to keep humans genuinely in the loop, not just nominally present. https://thesuperskills.com/research/superskill-curiosityhttps://thesuperskills.com/research/superskill-curiosityWed, 26 Aug 2026 09:00:00 GMTWhy curiosity is the foundational SuperSkill, how AI erodes it, and what it looks like when it is present or absent in people and organisations. https://thesuperskills.com/research/common-ai-transformation-challengeshttps://thesuperskills.com/research/common-ai-transformation-challengesWed, 26 Aug 2026 09:00:00 GMTThe recurring challenges leaders face with AI transformation, from tool overwhelm to automation anxiety, and how to map where AI adds value versus where human judgement matters. https://thesuperskills.com/research/superskill-change-readinesshttps://thesuperskills.com/research/superskill-change-readinessWed, 26 Aug 2026 09:00:00 GMTWhy change readiness is a baseline capability, not a crisis skill; how it differs from resilience; and how AI can both support and erode it. https://thesuperskills.com/research/superskill-big-picture-thinkinghttps://thesuperskills.com/research/superskill-big-picture-thinkingWed, 26 Aug 2026 09:00:00 GMTWhy the capacity to see systems, second-order effects and long-term patterns decides the quality of judgement, and how AI both extends and threatens it. https://thesuperskills.com/research/ai-transformation-not-change-mangement-problemhttps://thesuperskills.com/research/ai-transformation-not-change-mangement-problemWed, 26 Aug 2026 09:00:00 GMTChange management assumes a known destination, a plannable journey, and resistance as the obstacle. AI breaks all three. What replaces the plan is a quarterly practice. https://thesuperskills.com/research/ai-peoplehttps://thesuperskills.com/research/ai-peopleWed, 26 Aug 2026 09:00:00 GMTA structured reference to the people building, critiquing, governing and explaining AI. Curated by Rahim Hirji as a resource for professionals who need to know who shapes it. https://thesuperskills.com/research/will-ai-replace-my-jobhttps://thesuperskills.com/research/will-ai-replace-my-jobSun, 09 Aug 2026 09:00:00 GMTNot as a whole, but in parts, and which parts decides everything. The evidence on task exposure, the null effect on pay so far, and how to assess your own position in twenty minutes. https://thesuperskills.com/research/will-ai-replace-entry-level-jobshttps://thesuperskills.com/research/will-ai-replace-entry-level-jobsSun, 09 Aug 2026 09:00:00 GMTWill AI replace entry-level and graduate jobs? The evidence is that AI is dissolving junior tasks and pushing senior demands down into entry roles, not simply deleting jobs. The real risk is to the leadership pipeline. What the data shows, and what to do. https://thesuperskills.com/research/why-learn-to-prompt-is-weak-career-advicehttps://thesuperskills.com/research/why-learn-to-prompt-is-weak-career-adviceSun, 09 Aug 2026 09:00:00 GMTPrompting is real, but it is not scarce, not durable and not the constraint. What the evidence says about prompt skill, and the three capabilities that appreciate as prompting gets easier. https://thesuperskills.com/research/who-owns-verification-when-ai-does-the-workhttps://thesuperskills.com/research/who-owns-verification-when-ai-does-the-workSun, 09 Aug 2026 09:00:00 GMTIn most organisations, nobody. Verification is the stage everybody assumes is happening and almost nobody has assigned. Since August 2026 the EU AI Act presupposes there is somebody assigned. https://thesuperskills.com/research/what-we-know-about-ai-and-human-capabilityhttps://thesuperskills.com/research/what-we-know-about-ai-and-human-capabilitySun, 09 Aug 2026 09:00:00 GMTA living synthesis separating what we know from what we suspect from what we do not know yet. Nineteen claims in three bands, each graded by evidence strength and linked to its source. Reviewed quarterly. https://thesuperskills.com/research/what-stays-humanhttps://thesuperskills.com/research/what-stays-humanSun, 09 Aug 2026 09:00:00 GMTMachine responses already rate as more empathetic than doctors' and more creative than most writers'. What survives is not a list of tasks AI cannot do, but three things it cannot be: accountable, the one who noticed, and the source of the question. https://thesuperskills.com/research/what-should-a-board-ask-about-aihttps://thesuperskills.com/research/what-should-a-board-ask-about-aiSun, 09 Aug 2026 09:00:00 GMTTwelve questions, each paired with the answer that should worry you. Boards are asking the wrong questions and getting reassuring answers to them. https://thesuperskills.com/research/chro-guide-to-aihttps://thesuperskills.com/research/chro-guide-to-aiSun, 09 Aug 2026 09:00:00 GMTA CHRO guide to AI: the move from skills management to capability design, the EU AI Act Article 4 literacy obligation as amended in 2026, and the six things HR has to own. https://thesuperskills.com/research/outsourced-recognitionhttps://thesuperskills.com/research/outsourced-recognitionSun, 09 Aug 2026 09:00:00 GMTOutsourced recognition is what happens when the expression of noticing another person is delegated to AI, so the words arrive without the seeing that used to produce them. Definition, evidence and what to do. https://thesuperskills.com/research/what-is-meaningful-human-oversighthttps://thesuperskills.com/research/what-is-meaningful-human-oversightSun, 09 Aug 2026 09:00:00 GMTSince 2 August 2026, Article 14 of the EU AI Act sets out what oversight of a high-risk AI system must enable, and it names automation bias in the legislation. Far more demanding than a human reviewing output. https://thesuperskills.com/research/what-is-human-ai-collaborationhttps://thesuperskills.com/research/what-is-human-ai-collaborationSun, 09 Aug 2026 09:00:00 GMTThe honest definition starts with an inconvenient finding: a 2024 Nature Human Behaviour meta-analysis of 106 experiments found human-AI combinations performed worse on average than the better of human alone or AI alone. https://thesuperskills.com/research/what-is-cognitive-offloadinghttps://thesuperskills.com/research/what-is-cognitive-offloadingSun, 09 Aug 2026 09:00:00 GMTCognitive offloading is the use of an external tool to reduce the mental demand of a task. The definition, where the concept comes from, what the evidence shows about the trade it makes, and why it matters for AI. https://thesuperskills.com/research/capability-debthttps://thesuperskills.com/research/capability-debtSun, 09 Aug 2026 09:00:00 GMTCapability debt is the accumulated loss of human knowledge, skill and judgement that builds up when an organisation automates work faster than it redesigns how people learn through doing. Definition, causes, evidence and how to reduce it. https://thesuperskills.com/research/what-is-automation-biashttps://thesuperskills.com/research/what-is-automation-biasSun, 09 Aug 2026 09:00:00 GMTAutomation bias is the tendency to accept automated output without proper scrutiny. It appears in experts as readily as novices, resists training, and splits into commission and omission errors, only one of which anyone designs against. https://thesuperskills.com/research/what-does-ai-literacy-mean-for-leadershttps://thesuperskills.com/research/what-does-ai-literacy-mean-for-leadersSun, 09 Aug 2026 09:00:00 GMTArticle 4 of the EU AI Act has required AI literacy since February 2025, and enforcement began in August 2026. It applies at every risk tier. Most organisations have bought tool training and called it literacy. https://thesuperskills.com/research/predictionshttps://thesuperskills.com/research/predictionsSun, 09 Aug 2026 09:00:00 GMTRahim Hirji has published an annual reading of emerging technology and human capability every year since 2017. A dated public record, including the 2019 list that flagged soft skills as the workplace X-factor and human-plus-machine 'Iron Men' years before it was obvious. https://thesuperskills.com/research/evidencehttps://thesuperskills.com/research/evidenceSun, 09 Aug 2026 09:00:00 GMTA curated, continually updated reference to the best research on AI and human capability, from the World Economic Forum, PwC, EY, MIT, the NBER, Harvard and BCG and the key academic studies. Each entry gives the finding, what it shows and what it does not prove, with a link to the original. https://thesuperskills.com/research/the-best-writing-on-aihttps://thesuperskills.com/research/the-best-writing-on-aiSun, 09 Aug 2026 09:00:00 GMTA chronology from March 2023 to August 2026. The stories got louder faster than the labour market changed, and the question migrated from what will AI do to the economy to what will living with AI do to people. https://thesuperskills.com/research/timelinehttps://thesuperskills.com/research/timelineSun, 09 Aug 2026 09:00:00 GMTBox of Amazing has published weekly since January 2017, five years and ten months before ChatGPT. The dated record: automation and employment in May 2017, soft skills in December 2018, AI tutors in September 2020. https://thesuperskills.com/research/canonhttps://thesuperskills.com/research/canonSun, 09 Aug 2026 09:00:00 GMTA classified map of the field: 84 works on what increasingly capable AI does to human capability, judgement, learning and expertise. Each with a finding, an honest evidence-strength note and a reading. Inclusion does not mean endorsement. https://thesuperskills.com/research/delegation-boundary-maphttps://thesuperskills.com/research/delegation-boundary-mapSun, 09 Aug 2026 09:00:00 GMT"Keep a human in the loop" does not say which human, at which point, on what grounds. Nine stages of work, and for each one: what the human retains, what AI may assist with, what AI may execute, and who owns the outcome. https://thesuperskills.com/research/human-skills-in-the-age-of-aihttps://thesuperskills.com/research/human-skills-in-the-age-of-aiSun, 09 Aug 2026 09:00:00 GMTWhich human skills matter most as AI improves, and why 'AI-proof jobs' is the wrong frame. What AI can assist, what it can only imitate, and the capabilities that still require human accountability, mapped to the seven SuperSkills. https://thesuperskills.com/research/human-at-the-starthttps://thesuperskills.com/research/human-at-the-startSun, 09 Aug 2026 09:00:00 GMTWhere does accountability live when a machine helps decide? Rahim Hirji's three positions, Human at the Start (HATS), Human in the Loop, and Human at the End (HATE), why 'we keep a human in the loop' protects nobody on its own, and the one test that restores accountability. https://thesuperskills.com/research/how-this-research-workshttps://thesuperskills.com/research/how-this-research-worksSun, 09 Aug 2026 09:00:00 GMTThe method: how sources are selected and graded, how contradictory evidence is handled, where AI is used and where it is not, how terms are attributed, how corrections work, and what interests are disclosed. https://thesuperskills.com/research/how-should-leaders-respond-to-aihttps://thesuperskills.com/research/how-should-leaders-respond-to-aiSun, 09 Aug 2026 09:00:00 GMTAdoption ran faster than the PC, measured pay effects are near zero, and the structure of work is already moving. The four postures organisations take, and the six decisions a leadership team should actually make. https://thesuperskills.com/research/human-ai-decision-makinghttps://thesuperskills.com/research/human-ai-decision-makingSun, 09 Aug 2026 09:00:00 GMTAcross 106 studies, human-AI combinations performed worse than the better of human or AI alone, with losses concentrated in decision-making. Why human in the loop is not a design, and how to allocate decisions properly. https://thesuperskills.com/research/ai-workforce-strategyhttps://thesuperskills.com/research/ai-workforce-strategySun, 09 Aug 2026 09:00:00 GMTAdoption has been faster than the internet and the earnings data has not moved. The evidence on AI adoption, task reorganisation and workforce capability, and the four decisions a real workforce strategy makes. https://thesuperskills.com/research/how-to-assess-students-when-ai-can-do-the-assignmenthttps://thesuperskills.com/research/how-to-assess-students-when-ai-can-do-the-assignmentSun, 09 Aug 2026 09:00:00 GMTYou cannot do it by detection, and Sydney has admitted prohibition is unenforceable. Denmark now requires oral defence of every home-written exam. What the evidence says, and the two measurements every institution should separate. https://thesuperskills.com/research/how-humans-learn-with-aihttps://thesuperskills.com/research/how-humans-learn-with-aiSun, 09 Aug 2026 09:00:00 GMTAI raises performance while the tool is present and can lower capability once it is gone. The evidence on learning, deliberate practice and generative AI, and how to protect the reps that build judgement. https://thesuperskills.com/research/using-ai-without-dependencyhttps://thesuperskills.com/research/using-ai-without-dependencySun, 09 Aug 2026 09:00:00 GMTDependency is not heavy use. It is use that has replaced a capability you no longer hold. The evidence on cognitive offloading and over-reliance, a five-question self-check, and the practice that keeps AI as leverage. https://thesuperskills.com/research/staying-valuable-in-the-age-of-aihttps://thesuperskills.com/research/staying-valuable-in-the-age-of-aiSun, 09 Aug 2026 09:00:00 GMTYour job is not disappearing, it is being unbundled, and value moves to whichever tasks still require expertise. The economic evidence on which automation raises wages and which lowers them, and what it means for your career. https://thesuperskills.com/research/how-do-i-know-when-ai-is-wronghttps://thesuperskills.com/research/how-do-i-know-when-ai-is-wrongSun, 09 Aug 2026 09:00:00 GMTYou cannot tell from the output. A model's confidence is a property of its writing style, not its knowledge. The only reliable signal is a frontier map: knowing which categories of question it handles badly in your own domain. https://thesuperskills.com/research/ai-and-critical-thinkinghttps://thesuperskills.com/research/ai-and-critical-thinkingSun, 09 Aug 2026 09:00:00 GMTDoes using AI or ChatGPT weaken critical thinking? The evidence, the uncertainty, and the SuperSkills view: critical thinking does not disappear, it migrates to verification and thins unless you design against it. https://thesuperskills.com/research/ai-and-human-judgementhttps://thesuperskills.com/research/ai-and-human-judgementSun, 09 Aug 2026 09:00:00 GMTAI can improve the quality of an output while weakening the capability that would have produced it. The evidence on AI, critical thinking and human judgement, and what leaders should do, from the research behind SuperSkills. https://thesuperskills.com/research/does-ai-make-everyone-think-alikehttps://thesuperskills.com/research/does-ai-make-everyone-think-alikeSun, 09 Aug 2026 09:00:00 GMTYes, and not by making anyone worse. In a 2024 Science Advances experiment, AI-assisted stories were rated more creative and were markedly more similar to each other. Individually better, collectively narrower. https://thesuperskills.com/research/correctionshttps://thesuperskills.com/research/correctionsSun, 09 Aug 2026 09:00:00 GMTA public log of everything this research got wrong, with the original claim, what was found, and what changed. Including four coinage claims withdrawn and one term where independent prior use was found. https://thesuperskills.com/research/ai-agents-and-human-judgementhttps://thesuperskills.com/research/ai-agents-and-human-judgementSun, 09 Aug 2026 09:00:00 GMTWhen AI agents can plan, act and make recommendations, the human's job shifts from doing to deciding what to delegate, setting the boundaries and owning the outcome. The delegation boundary map, human in the loop versus human on the hook, and who is accountable when an agent is wrong. https://thesuperskills.com/research/questionshttps://thesuperskills.com/research/questionsSun, 09 Aug 2026 09:00:00 GMTAn evolving map of the questions that matter as AI becomes more capable. 180 real questions across twelve territories, marked answered, partial or open. The open ones stay visible on purpose.