A Conversation for Humanity — Deverout Graham
We talk. We listen. We look. We learn.
The question
If artificial intelligence helps people complete work faster, who should receive the time and value it creates?
This is not only a question for technology companies. It concerns workers, employers, customers, governments and countries still trying to build productive economies. The answer will shape whether AI becomes a tool for shared progress or another source of distance between those who own systems and those whose work changes around them.
Perspective one: let productivity grow first
Businesses take real risks when they invest in AI: money, management attention, training and changes to established workflows. If an organisation cannot keep enough of the return to justify that effort, it may not invest at all. Higher productivity can also support lower prices, better services, new products and jobs that did not previously exist. On this view, the first responsibility is to make adoption work; wider benefits can follow as firms grow and competition spreads them.
There is a sound objection to promising a large “AI dividend” before it has been measured. The International Labour Organization’s 2026 review found that productivity gains from generative AI are real in some settings but uneven and often not yet verified. Worker-reported time savings of a few per cent of working hours had not yet translated into higher measured output, earnings or employment. If the gains are uncertain, a rigid sharing formula could burden smaller firms before the benefits are clear.
Perspective two: make sharing part of the design
Waiting for benefits to “trickle down” can leave workers with more monitoring, faster workloads or less secure employment while others capture the savings. Sharing does not have to mean dividing every pound equally. It can mean better wages, paid training, safer workloads, shorter hours where feasible, improved public services, or workers having a voice in how new systems change their jobs. Collective bargaining and social dialogue can make those choices visible before new routines become difficult to reverse.
The objection is practical: firms cannot distribute gains they have not actually made. AI may save time on one task while adding verification, integration and oversight elsewhere. A benefit that exists in a demonstration may disappear when the whole workflow is counted. Sharing promises should therefore follow transparent measurement, not impressive claims from a pilot or a vendor.
What the evidence says—and does not say
The ILO’s May 2026 review of the “aggregation paradox” reports that task-level productivity gains can be substantial, while firm-level findings are mixed and adoption remains uneven. It finds no clear AI-driven productivity growth in economy-wide statistics so far. The review argues that broad gains depend on diffusion, complementary investment, workplace redesign, skills and institutions that shape how benefits are shared.
That evidence does not prove that AI will fail to raise productivity, or that workers must lose. It tells us that the outcome is still being made. We should be careful with both the promise of automatic prosperity and the prediction of inevitable displacement. The honest measure is not how quickly a tool produces a first answer, but what happens to quality, total work, earnings, jobs and human judgement over time.
A constructive next step
For one real workplace adopting AI, compare the old and new process over a defined period. Measure total time, rework, quality, customer outcomes, workload and earnings—not just the minutes saved on one task. Then ask workers and managers together how any verified gain should be used. That gives the next conversation something firmer than optimism or fear: shared evidence.
Which should come first: proving that AI creates a real productivity gain, or agreeing in advance how workers will have a voice in sharing it? Perhaps the answer is to do both—measure openly, then negotiate fairly.
Join the conversation: We talk. We listen. We look. We learn.
Sources
- International Labour Organization, “The impact of GenAI on jobs, productivity and work organization: a review of the empirical evidence”, 1 June 2026.
- International Labour Organization, “The Aggregation Paradox of AI”, 6 May 2026.
Deverout Graham is Managing Partner of Deverout & Associates, Bedford · London · Global Affiliates. Visit deuerout.com.
