A friend's reflection on ADHD, artificial intelligence and a split-second intervention in a supermarket raises a larger question. As technology expands what people can do, how will our institutions decide whom to trust with that power?
“If enough of us decide to feel something again ... we have a very good chance to change the world for the better.”
Food for thought
This article was prompted by a post from my friend Nat Schooler, ‘We Have Stopped Feeling. That Is the Real Emergency Not AI!’. His reflection moved across ADHD, artificial intelligence, learning and a split-second intervention in a supermarket before asking whether the greater danger lies in AI itself or in a weakening of human care and responsibility. I take that post as a stimulus rather than a theory to be defended.
Not every causal connection in the post can carry the weight placed on it. A friendship network cannot establish a population-level pattern, and isolated incidents do not sustain broad claims about competence, economics or social behaviour. Yet the post surfaces three propositions worth separating: cognitive difference may become more valuable in an AI-enabled world; people must learn to use the tools now available to them; and increased capability without commensurate judgement may make us more efficient without making us better.
The third proposition is the one I want to pursue. The skills question matters, but it is no longer sufficient. Employers, universities and governments are already debating AI literacy, reskilling and productivity. The harder question begins after people have acquired the tools. What are we preparing them to do with their enlarged capability, and what qualities do our institutions notice and reward when they decide who should lead?
The pattern and its limits
“It stopped being a coincidence and started looking like a pattern.”
Nat may indeed have noticed a pattern. Neurodivergent people often recognise one another through ways of thinking that feel familiar: rapid association, intense curiosity, unusual pattern recognition, deep focus on compelling problems and impatience with conventions that seem to exist only because nobody has challenged them. People also build self-selecting networks. We are drawn towards those with whom conversation moves at the right speed, whose minds take similarly unexpected routes and who do not require constant translation. A pattern within a network is therefore a worthwhile hypothesis, but it is not evidence that ADHD causes success or that exceptional people are predominantly neurodivergent.
We should also resist the seductive move from cognitive difference to moral distinction. ADHD does not confer integrity, empathy or courage. Neurotypicality does not diminish them. Neurodivergence describes differences in cognition and neurological functioning, not a superior class of human beings. Recent workplace research shows both valuable strengths and serious structural barriers. Neurodivergent employees may contribute creativity, analytical ability and alternative approaches to problem solving, while still encountering stigma, masking, unsuitable recruitment processes and inconsistent support (Vargas-Salas et al., 2025).
The language of the ADHD superpower can be affirming after years of deficit-based description, but it can also become another demand to be exceptional. A person should not have to turn disability into commercial advantage before an organisation will make room for them. The strongest claim is more modest and more useful: cognitive variety can generate distinctive value when environments are designed to recognise it. That places responsibility on institutions as well as individuals.
When technology works with the mind
“Today we live in an age where just the right tools exist to work with how your brain actually operates instead of against it.”
This is the most persuasive part of Nat's argument. As someone with AuDHD, I recognise the relief of a tool that can hold the scaffolding of a task while my attention moves across its architecture. Generative AI can help externalise sequence, retrieve a half-formed connection, reorganise material, translate between formats and reduce the administrative drag between an idea and its expression. Used well, it can release energy that was previously spent compensating for systems built around a narrow model of attention and organisation.
The wider evidence supports the possibility of meaningful augmentation, although it does not justify technological triumphalism. In a study of 5,172 customer-support agents, access to a generative AI assistant increased productivity by 15 per cent on average. The largest gains went to less experienced and lower-skilled workers, suggesting that AI can distribute some forms of tacit knowledge more widely. The most experienced workers saw smaller gains and, in some cases, slight reductions in quality (Brynjolfsson, Li and Raymond, 2025).
Other research describes a jagged technological frontier. AI can improve speed and quality when a task falls within its effective range, while making performance worse when users trust it beyond that range (Dell'Acqua et al., 2023). A 2025 study of knowledge workers also found that greater confidence in generative AI was associated with less reported critical thinking, whereas confidence in one's own expertise was associated with more (Lee et al., 2025). These findings complicate any simple instruction to adopt the tools quickly. Effective use requires enough subject knowledge to recognise when the tool is wrong, enough humility to question an attractive answer and enough responsibility to remain accountable for the result.
AI can remove unproductive friction, but some friction is the work. Comparing evidence, sitting with ambiguity, noticing a contradiction and forming a judgement are not administrative obstacles to thinking. They are thinking. The educational and organisational challenge is to distinguish the burden that can safely be lifted from the cognitive effort that develops expertise.
Capability and competency are not calibre
I find it useful to separate three related dimensions. Within the developing IOWL architecture, Capability concerns what a person knows, understands and has the capacity to do. Competency concerns the effective application of that Capability to an appropriate standard. Calibre concerns the values-governed quality of judgement through which Capability and Competency are exercised, particularly when rules are incomplete, incentives are distorted, interests compete or consequences extend beyond the actor. A person may therefore be highly capable and demonstrably competent without necessarily exercising those assets with Calibre.
IOWL represents the interdependence of these dimensions through the Leadership Stool™: Capability, Competency and Calibre are separate but mutually necessary supports of dependable human performance. None can substitute for the others. AI is significant because it can extend Capability and produce the appearance of Competency at extraordinary speed without automatically strengthening the Calibre governing their use.
This is one expression of what IOWL describes as the Capacity Conundrum™: expanding human, technical and institutional capacity creates possibility, but it does not determine direction or guarantee conversion into trusted performance. The question is therefore not only what additional capacity AI creates, but what governs the purposes, judgements and consequences attached to its use.
Management scholarship has often treated integrity as a desirable but imprecise quality. Palanski and Yammarino (2007) sought to clarify it through consistency between words and actions and through conduct that remains coherent under pressure. Crossan et al. (2017) place judgement at the centre of a broader framework of leader character, connected to courage, accountability, humanity, humility, justice and integrity. This matters because character is not a decorative addition to technical performance. It shapes which goals people choose, which risks they notice, whose interests they consider and what they are prepared to do when nobody is rewarding them for it.
The distinction becomes more consequential as AI lowers the cost of producing fluent, plausible and polished work. When almost anyone can generate an impressive proposal, presentation or strategy in minutes, surface competence becomes a weaker signal. The scarce qualities move elsewhere: the knowledge needed to test an answer, the honesty to disclose uncertainty, the judgement to recognise harm, the courage to resist a convenient falsehood and the willingness to accept responsibility for a decision.
This concern is older than AI. Aristotle argued that virtue is formed through habituation rather than possessed as abstract knowledge. MacIntyre (2007) later distinguished the internal goods of a practice from external goods such as status, money and power, warning that institutions can corrupt the practices they are meant to sustain when external rewards dominate. Vallor (2016) applies virtue ethics to technological life and argues for the deliberate cultivation of qualities that help human beings live wisely amid rapid and uncertain change. AI has not invented the need for character. It has increased the reach of choices made without it.
This distinction also helps separate outside-in governance from inside-out judgement. Regulation, codes, policy and compliance can constrain conduct from the outside, but they cannot remove the need for human judgement when circumstances exceed the rulebook. Within IOWL, the Axiological Pyramid™ is used to explore that relationship between internal values and external expectations: formal controls matter, but they are not a substitute for the Calibre of the person exercising discretion.
What the supermarket incident really asks
Nat's supermarket story is vivid because it turns an abstract discussion into a moment of decision. He describes considering whether to record what was happening and deciding instead to intervene physically. I understand why that moment stayed with him. It contains a question about responsibility: when does an event occurring in public become something for which I, too, must decide what to do?
We should be careful, however, not to make physical intervention a simple test of moral worth. Entering a volatile situation can protect someone, or it can escalate danger. Recording can be passive spectacle, or it can preserve evidence. Calling for trained help may be wiser than acting alone. Research on the bystander effect is similarly more nuanced than the familiar story of crowds paralysed by indifference. A large meta-analysis found an overall reduction in helping when other bystanders were present, but the effect weakened in dangerous situations and changed when other people could provide physical support (Fischer et al., 2011).
The better lesson is that care requires judgement. Action is not automatically courageous because it is immediate, and caution is not automatically cowardice because it is quiet. The morally serious question is whether a person notices what is happening, accepts some responsibility for responding and chooses a proportionate course that considers the likely consequences.
Organisations rarely have to wait for a dramatic incident to observe this quality. It appears when somebody reports a safety concern before an accident; admits an error before it becomes a scandal; refuses to manipulate data; gives credit to a junior colleague; protects a person with less institutional power; challenges a profitable but harmful decision; or tells a senior group what it needs to hear rather than what will preserve the speaker's status. Most integrity is undramatic. That may be one reason institutions are so poor at recognising it.
Why institutions reward the wrong signals
Technical competence is easier to count than judgement. Confidence is easier to observe than humility. Individual output is easier to attribute than harm prevented through careful stewardship. Recruitment and promotion systems therefore rely heavily on proxies: polished applications, fluent interviews, visible busyness, short-term targets and the ability to narrate one's own impact. Those proxies have always favoured some personalities and backgrounds over others. AI now makes several of them cheaper to manufacture.
This does not make expertise irrelevant. It makes evidence of expertise more important and performance of expertise less reliable. A perfectly written application may reveal little about how a candidate reasons. A values statement may reveal little about what an organisation tolerates from its highest earner. A leader may speak eloquently about care while repeatedly externalising the cost of their decisions onto less powerful people. Integrity becomes visible only when words, actions, incentives and consequences are examined together.
Organisations also send powerful messages through exceptions. If a high performer is protected after behaviour that would end another person's career, the real value system is clear. If mentoring, prevention, dissent and care are praised but carry no weight in workload, recognition or promotion, employees learn that these activities are optional. Culture is produced by what institutions reward, excuse and repeatedly fail to notice.
This is where the claim that caring has become rare needs refinement. People who care may not be rare. Systems that can reliably identify, sustain and reward care are rare. Many people learn to conceal moral concern because it slows a target, complicates a decision or threatens a hierarchy. Others burn out from carrying responsibilities that the institution celebrates rhetorically but refuses to resource.
IOWL describes part of this problem as an HR Blind Spot™: institutions have developed increasingly sophisticated ways to verify qualifications, experience, knowledge and skills, yet remain far less systematic about developing and evidencing how those assets are likely to be exercised when judgement, trust, competing interests and consequence matter. AI widens that blind spot because polished outputs, fluent applications and rehearsed performances of expertise become easier to manufacture.
Developing and recognising calibre
At the Institute of One World Leadership, this is the problem that interests me most: how to develop, evidence and recognise Calibre, integrity and positive values with the seriousness routinely given to technical skill. IOWL's developing contention is not that Calibre can or should be reduced to a moral score. It is that the values-governed quality of judgement can be developed and evidenced more deliberately than institutions currently attempt. Positive Value Leadership provides one developmental architecture within that work, while the wider IOWL institutional architecture connects development with Human Calibre, formal recognition, integrity, employability, productivity and performance. Any credible approach must remain behavioural, contextual, evidence-based and open to challenge.
A practical agenda would include six changes.
- Define the conduct that mattersOrganisations should translate values into observable choices. Integrity may include disclosing uncertainty, keeping commitments, escalating material risks, resisting conflicts of interest, correcting errors and using authority consistently. Care may include noticing the effects of decisions on people with less power, not simply displaying warmth.
- Use evidence from more than one settingSelf-description is weak evidence of character. Selection and promotion should combine structured ethical scenarios, work samples, critical-incident interviews, carefully designed references, multi-source feedback and a record of decisions over time. The purpose is not to find moral perfection, but to understand how a person reasons and behaves when interests collide.
- Develop judgement through practicePeople learn character through repeated action, reflection and feedback. Ethical dilemmas, simulations, case discussion, coaching, mentoring and responsibility for real consequences can make judgement explicit. Development should include occasions when speaking up is difficult, information is incomplete and every available option carries a cost.
- Reward stewardship as real performancePromotion and recognition criteria should give weight to mentoring, prevention, collaboration, responsible challenge, long-term trust and the quality of the environment a person creates for others. Institutions should also remove the protective halo around technically brilliant people whose conduct repeatedly damages colleagues or public trust.
- Design for cognitive differenceNeuroinclusive work requires flexibility, clear communication, suitable sensory conditions, task customisation and routes to demonstrate competence that do not depend on a single social style. These conditions should enable contribution without turning diagnosis into a shortcut for presumed creativity, goodness or fit.
- Teach AI literacy with moral and epistemic literacyPeople need to know how to use AI, but also how to verify claims, protect data, disclose assistance, test bias, recognise uncertainty, preserve human accountability and decide when a task should not be delegated. NIST and UNESCO both place trustworthiness, responsibility and human oversight within serious AI governance rather than treating adoption as a purely technical matter (NIST, 2023; UNESCO, 2021).
Development also cannot end at a course, recruitment event or promotion decision. IOWL's emerging concept of Perpetual Calibre Development™ treats Calibre as something that should be revisited as role, authority, context, pressure and consequence change. Hindsight, insight and foresight become part of continual recalibration rather than a one-off judgement of character.
Recognition without moral surveillance
There is an obvious danger in trying to assess character. Values frameworks can become instruments of conformity. Employees who communicate differently may be marked down for style; cultural expectations may be presented as universal virtues; and people skilled in impression management may learn to perform the approved language. The answer is not to abandon recognition, but to make it more disciplined.
Evidence should concern conduct relevant to a role, not personality preference. Expectations should be transparent. Decisions should draw on multiple sources and allow challenge or appeal. Neurodivergence should never be inferred from behaviour or used as a proxy for either talent or Calibre. Nor should cultural style, confidence, fluency or conformity to a preferred leadership personality be allowed to stand in for values-governed judgement. Organisations should examine outcomes and patterns over time, including who bears the consequences of a leader's success.
Calibre also includes the capacity to revise. A person who never admits error may look consistent while lacking integrity. Learning, apology and repair matter because good judgement is not infallibility. The relevant question is whether someone can face evidence that threatens their preferred account, take responsibility without transferring blame and change their conduct.
The choice is not AI or humanity
“Are you still scared of AI or more scared of the loss of humanity?”
I would resist the choice. Concern about AI may itself express care for human dignity, employment, privacy, truth, safety and the distribution of power. It is possible to welcome tools that expand access and creativity while demanding limits, accountability and public scrutiny. The opposite of fear is not uncritical adoption. It is informed agency.
AI systems do not enter neutral environments. They enter organisations with existing incentives, inequalities and habits. A system that rewards volume will use AI to produce more volume. A system that rewards surveillance will use it to watch more people. A system committed to access may use it to remove avoidable barriers. A system that values professional judgement may use it to support decisions while retaining clear human accountability. The technology amplifies the purpose and discipline surrounding it.
This makes Calibre a performance question as well as an ethical one. The same technology can accelerate stewardship or accelerate error, distortion and harm. IOWL uses the term Calibre Risk™ for the exposure created when significant Capability or authority is exercised without commensurate values-governed judgement. As AI increases reach and speed, that exposure can increase even where technical performance appears to improve.
For that reason, the future of work cannot be divided into people who learned AI and people who failed to adapt. The more consequential divide may be between institutions that treat human beings as interchangeable operators of efficient systems and those that cultivate knowledge, judgement, responsibility and trust. The first may move quickly. The second has a better chance of knowing where it is going.
The question beneath the technology
The discussion that prompted this article matters because it refuses to separate technological change from moral life. AI is not only changing the tasks people perform. It is exposing what our institutions have chosen to value and how weak many of our methods remain for developing, evidencing and recognising the people we should trust with expanding capability.
Learning the tools is necessary. It will not be enough. We must become better at distinguishing fluent output from knowledge, confidence from judgement, visible action from wise action and declared values from integrity lived under pressure. We must also stop asking neurodivergent people to prove their worth through exceptional performance while leaving the environments that exhaust or exclude them unchanged.
The task ahead is to build institutions in which calibre can be developed, observed and consequentially rewarded; in which care does not depend on individual heroism; and in which powerful tools remain answerable to human purposes. If AI made everyone in an organisation faster tomorrow, what in that organisation would prevent faster carelessness? What would help people exercise better judgement? And who would it choose to recognise when the right action was quiet, inconvenient and costly?
Those questions reach further than fear. They ask what kind of people, leaders and institutions we intend to become while our capabilities are increasing.
That is why I now see the issue as larger than whether AI will make us more capable. The more demanding question is whether our institutions can strengthen the Calibre with which growing Capability is exercised — and recognise it when it appears in forms that are quiet, inconvenient, difficult to measure or costly to the person who acts with integrity.
References
- Aristotle (2009) The Nicomachean Ethics. Translated by D. Ross and revised by L. Brown. Oxford: Oxford University Press.
- Brynjolfsson, E., Li, D. and Raymond, L.R. (2025) 'Generative AI at Work', The Quarterly Journal of Economics, 140(2), pp. 889-942. https://doi.org/10.1093/qje/qjae044
- Crossan, M.M., Byrne, A., Seijts, G.H., Reno, M., Monzani, L. and Gandz, J. (2017) 'Toward a Framework of Leader Character in Organizations', Journal of Management Studies, 54(7), pp. 986-1018. https://doi.org/10.1111/joms.12254
- Dell'Acqua, F. et al. (2023) 'Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality', Harvard Business School Working Paper 24-013.
- Fischer, P. et al. (2011) 'The Bystander-Effect: A Meta-Analytic Review on Bystander Intervention in Dangerous and Non-Dangerous Emergencies', Psychological Bulletin, 137(4), pp. 517-537. https://doi.org/10.1037/a0023304
- Fore, P. (2026) 'Palantir's billionaire CEO says only two kinds of people will succeed in the AI era', Fortune, 24 March.
- Lee, H-P., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R. and Wilson, N. (2025) 'The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers', Proceedings of CHI 2025, Article 1121. https://doi.org/10.1145/3706598.3713778
- MacIntyre, A. (2007) After Virtue: A Study in Moral Theory. 3rd edn. Notre Dame, IN: University of Notre Dame Press.
- National Institute of Standards and Technology (NIST) (2023) Artificial Intelligence Risk Management Framework AI RMF 1.0. Gaithersburg, MD: NIST. https://doi.org/10.6028/NIST.AI.100-1
- Schooler, N. (2026) 'We Have Stopped Feeling. That Is the Real Emergency Not AI!', Future Proofed Leader, 17 September. Available online (Accessed: 17 September 2026).
- Palanski, M.E. and Yammarino, F.J. (2007) 'Integrity and Leadership: Clearing the Conceptual Confusion', European Management Journal, 25(3), pp. 171-184. https://doi.org/10.1016/j.emj.2007.04.006
- UNESCO (2021) Recommendation on the Ethics of Artificial Intelligence. Paris: UNESCO.
- Vallor, S. (2016) Technology and the Virtues: A Philosophical Guide to a Future Worth Wanting. Oxford: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780190498511.001.0001
- Vargas-Salas, O., Alcazar-Gonzales, J.C., Fernandez-Fernandez, F.A., Molina-Rodriguez, F.N., Paredes-Velazco, R. and Carcausto-Zea, M.L. (2025) 'Neurodivergence and the Workplace: A Systematic Review of the Literature', Journal of Vocational Rehabilitation, 63(1), pp. 83-94. https://doi.org/10.1177/10522263251337564