AI can move the productivity bottleneck rather than remove it: IOWL raises Human Calibre at OECD Global Forum

BELFAST, 6 October 2026. The Institute of One World Leadership (IOWL) has raised a deeper productivity question following an exchange during the 2026 Annual Conference of the OECD Global Forum on Productivity in San José, Costa Rica: if artificial intelligence increases what workers can produce but the judgement required to validate, challenge and responsibly use that output is pushed upwards to managers, has the productivity constraint been removed or merely moved?

The question emerged during Session 6, AI, Skills, and Addressing Emerging Skill Gaps in the Age of AI, after IOWL President and Executive Director Professor Noel Ferguson asked whether policy risks concentrating on the part of human contribution it can already measure, principally skills, while leaving the quality of judgement, responsibility and decision-making insufficiently distinguished inside multifactor productivity.

Professor Ferguson asked whether that under-examined dimension, which IOWL describes as Human Calibre, could become more rather than less important as AI increasingly democratises Capability.

From Calibre to management: a revealing reframing

The question was initially interpreted in the session as being about management and organisational skills. In response, MIT researcher Neil Thompson suggested that more people may increasingly resemble managers as they supervise AI agents, making sure those agents are doing the right things and knowing how to check their work.

IOWL believes that observation is important, but that it also exposes the issue the Calibre proposition is intended to address.

“If every AI-assisted employee requires a manager to supply the judgement, responsibility and quality control that the employee does not exercise themselves, AI has not removed a productivity constraint. It has moved the constraint up the organisation.”
Professor Noel Ferguson, President & Executive Director, IOWL

IOWL argues that the productivity opportunity is not simply to make managers better at checking AI-assisted work. It is to develop sufficient Calibre and integrous agency throughout the workforce so that people can exercise powerful new Capability responsibly at the point where it is used.

The quality-review burden is already visible

Earlier in the same session, Thompson explicitly identified several activities associated with AI adoption as organisational overheads, including AI integration, ethics and compliance, and AI quality review. He observed that organisations can spend substantial time on the activities required to make AI systems work effectively, limiting their ability to capture the productivity benefit.

For IOWL, this creates a direct bridge between the conference evidence and the Calibre proposition.

AI can dramatically increase the amount of material an individual is able to generate, analyse or process. That is an increase in Capability. But if the individual cannot determine whether the output is accurate, appropriate, responsible or suitable for action, the apparent gain can create a downstream requirement for checking, correction, escalation and re-work.

The organisation may therefore produce more output while simultaneously creating more verification work.

“The critical productivity question is not only how much AI helps somebody produce. It is how much additional checking, correction and escalation that output creates elsewhere in the organisation.”

“Capability can scale very quickly. If Calibre does not scale with it, the bottleneck moves from production to judgement.”
Professor Noel Ferguson, President & Executive Director, IOWL

A Calibre Burden-Shift™

IOWL describes the mechanism as a potential Calibre Burden-Shift™.

Where AI Capability is distributed without corresponding human Calibre, the burden of judgement does not disappear. It moves to supervisors, managers, compliance teams, quality-assurance functions or other expert reviewers.

That matters because management capacity is itself scarce. A manager who must repeatedly verify work that should have been responsibly evaluated at source has less time for genuinely managerial activity, including direction, development, coordination, resource allocation and strategic judgement.

The potential productivity chain can therefore work in two different ways:

AI Capability rises + insufficient Calibre = verification, re-work and escalation rise = management burden rises = realised productivity gain is reduced.

AI Capability rises + sufficient Calibre and integrous agency = responsible judgement at the point of use rises = avoidable verification and re-work fall = more of the potential productivity gain is realised.

IOWL presents this as a research proposition, not as an already established causal equation. The Institute argues that the mechanism is sufficiently plausible and economically consequential to warrant empirical investigation.

Judgement is not simply prediction

The exchange also raised a further boundary question. Thompson suggested that many questions commonly described as judgement may be reframed as prediction problems that AI can solve well.

IOWL accepts that many decisions contain predictive components. It argues, however, that prediction and judgement should not be treated as synonymous.

An AI system may predict which course of action is most likely to maximise a measurable return. That prediction does not determine whether the action should be taken, whether its consequences for others are acceptable, whether material information should be disclosed, whether an apparently optimal recommendation should be challenged, or who accepts responsibility for the final decision.

Those are questions of agency as well as analysis.

“Prediction can tell us what is likely to happen. Capability can tell us what can be done. Neither, by itself, determines what should be done or the quality with which the person responsible exercises that choice.”
Professor Noel Ferguson

Macro productivity, micro causes

IOWL believes the discussion also highlights a level-of-analysis problem in productivity research.

Productivity is commonly examined at economy, sector and firm level. Yet productive performance emerges from activity inside organisations, including thousands of individual decisions about resources, priorities, risk, responsibility, people and the use of available technology.

At this micro level, employers encounter the effects of weak judgement, poor responsibility, value misalignment, inappropriate behaviour, avoidable re-work and the misuse of otherwise considerable Capability.

Current HR evidence makes the issue particularly relevant. SHRM's 2026 Talent Trends research reports that 80 per cent of HR professionals experiencing recruitment difficulty identify systems and resource-management capabilities, including judgement, decision-making, complex problem-solving and time management, as particularly difficult to find. Separate SHRM benchmarking places critical thinking, judgement and decision-making among the highest workforce-development priorities for the next five years.

IOWL argues that these qualities should not automatically be absorbed into an ever-expanding definition of skills without first asking whether an important qualitative dimension is being obscured.

“Economists understandably look at productivity from the top down. Employers experience many of its causes from the bottom up.”

“Organisations can recruit people because they have demonstrated the required skills and Capability, then discover that performance problems arise not because those people cannot do the job, but because of the quality with which they exercise that Capability.”
Professor Noel Ferguson

The productivity question becomes more precise

IOWL's proposition is therefore no longer simply that Calibre may matter to productivity. The emerging research question is more specific:

As AI democratises Capability, do differences in Human Calibre determine how much of the potential productivity gain is realised before it is absorbed by verification, correction, re-work, escalation and management overhead?

A related question is whether repeated micro-level differences in judgement, responsibility, integrity, discernment and the quality with which Capability is exercised aggregate into economically meaningful differences in firm, sector and economy-wide performance.

IOWL is not claiming that Human Calibre has already been demonstrated to constitute a component of multifactor productivity. It is proposing that the relationship now warrants interdisciplinary investigation involving economists, employers, HR leaders, universities and researchers working on human-AI interaction.

From architecture to a testable productivity proposition

The intervention follows IOWL's 29 September announcement that it is opening its Calibre architecture and Calibre Credits™ to universities and employers.

That architecture is intended to make the distinction between Capability and Calibre operational rather than rhetorical. IOWL distinguishes Capability, what a person knows or is able to understand; Competency, what a person can demonstrably do; and Calibre, the quality with which that Capability and Competency are exercised.

Its wider development architecture includes Positive Value Leadership™, Reflective Transformation, assessed Calibre development, Calibre Credits™, Statements of Professional Standing, the Passport of Probity™ and Perpetual Calibre Development™.

Professor Ferguson added:

“AI is not making the human question disappear. It is relocating it.”

“When production becomes easier, judgement becomes the scarce point. If that judgement exists only at the top of the organisation, management becomes the bottleneck. If responsible, integrous agency is developed throughout the organisation, AI has a much greater chance of delivering genuine productivity rather than simply more output.”

“In the age of AI, productivity needs more than Capability. It needs Calibre close to the point where Capability is exercised.”

Read the 29 September announcement: Productivity needs more than Capability in the age of AI >

Explore Calibre & Productivity >

Participation in OECD forums does not imply OECD endorsement, validation or adoption of IOWL's frameworks or propositions.

Ends.