IOWL Calibre Briefing - editorial working copy
AI is increasing human capability. That may make judgement more valuable, not less
Generative AI can make sophisticated forms of capability available to far more people at far greater speed. The technology debate therefore has a human question at its centre: who decides what should be done, what should be trusted and what consequences are acceptable?
The most obvious effect of generative AI is an expansion of capability. Tasks that once demanded specialist knowledge or substantial time can increasingly be accelerated by systems that draft, analyse, code, translate, summarise and generate. Governments and employers are pursuing that potential because it may improve productivity, services and innovation.
But an increase in capability does not settle the direction in which that capability will be used. That is the connection between today's AI debate and a line of thinking that began at the Institute of One World Leadership in 2013, long before the current generative-AI wave. IOWL's original question was already about the gap between what people can do and what governs how they choose to do it.
The institute calls that gap the Capacity Conundrum. More education, skill, technology or authority increases the range of possible action. It does not determine which action will be chosen. AI simply makes the conundrum move faster because it can expand the scale and speed of individual and organisational capacity.
The human oversight problem illustrates the point. It is easy to say that a person should remain in the loop. It is harder to ask what qualities make that oversight meaningful. Does the person understand enough to challenge the output? Are they willing to do so under pressure? Can they recognise manipulated or incomplete information? Do incentives reward caution or speed? Who accepts responsibility when the machine's answer is convenient but questionable?
International AI principles already recognise the importance of human agency and oversight. IOWL's contribution is not a competing technical safety model. It focuses on the Calibre of the human actor: the quality of judgement, values and agency brought to the use of increasingly powerful systems.
That framing also changes the employment story. If AI narrows some capability gaps by giving more people access to similar analytical or creative tools, the differentiating value of judgement may rise. Employers may care more about who can question an answer, disclose uncertainty, protect confidential information, understand consequence, resist convenient fabrication and use powerful tools in ways aligned with organisational and societal responsibility.
Positive Value Leadership gives IOWL a language for that development. Its six Exemplars include Probity, People, Perception and Participation, bringing together integrity, information awareness, questioning, humanity, collaboration and other defined Positive Values. The claim is not that a list of values solves AI governance. The proposition is that human responsibility cannot be engineered entirely outside the human being.
AI also magnifies the downside of poor judgement. A weak decision that once affected one document, one customer or one process can now be replicated at extraordinary speed. Automation can scale good practice, but it can also scale error, bias, manipulation or misplaced confidence. The more powerful the capability, the greater the potential consequence of using it badly.
This creates an unusual policy tension. Organisations are being encouraged to adopt AI quickly while simultaneously being told to maintain human oversight and trustworthy governance. The quality of that human layer may become the limiting factor. Tools can be deployed faster than judgement can be developed.
The question IOWL puts into that debate is therefore intentionally simple. AI is increasing what people and institutions can do. Are we investing with equal seriousness in the Calibre that governs what they should do, why they should do it and who will live with the consequences?