40% of jobs will be "affected" by AI. The verb is doing a lot of work.
The IMF says AI will touch 40% of jobs worldwide — and that half the exposed roles may gain, not lose. Which half your people land in is a leadership choice.
⚖️ 40% of jobs will be "affected" by AI. The verb is doing a lot of work.
The IMF opened Davos week with a number built for headlines: AI is likely to affect nearly 40% of jobs worldwide, and about 60% in advanced economies. Kristalina Georgieva's framing was appropriately sober — in most scenarios, inequality worsens unless policymakers act.
But the detail worth sitting with is buried one paragraph down: of the exposed jobs, roughly half may benefit — AI enhancing the work, lifting productivity and potentially incomes — while the other half see tasks displaced. Same exposure, opposite outcomes.
Which raises the question the headline skips: what decides which half a given worker lands in?
At the national level, that is policy. But inside an organisation — where most of us actually operate — I'd argue it is decided by choices leaders are making right now, mostly without noticing:
Whether AI is deployed to assist the person or to replace the task list. The same technology, pointed at "help our people do this better" versus "do this so we need fewer people," produces the two halves of the IMF's split.
Who gets access and fluency first. If AI capability concentrates in a small technical group, you have built the inequality dynamic inside your own walls — the augmented few, the exposed many. Broad access with real enablement is the counter-programme.
Whether anyone redesigns the role. A job with three tasks automated and nothing added is a shrinking job. The organisations that come out of this well will treat automation as freed capacity to reinvest — in judgment, relationships, harder problems — not as headcount arithmetic.
There is also a workforce-trust dimension that reports like this create whether we like it or not. Your employees will read the "40%" headline. If leadership says nothing, they will fill the silence with the pessimistic interpretation — and people who believe AI is being brought in to replace them do not adopt it, do not improve it, and quietly work against it. The augmentation story has to be told explicitly, and then — harder — it has to be true.
A question for every leadership team reading the IMF's number: for your ten most common roles, could you say specifically what AI will take off each plate, and what you intend to put on it instead? Until that second column is filled in, "affected" is just a number waiting for someone else to choose its meaning.
Reading: https://www.cnbc.com/2024/01/15/imf-warns-ai-to-hit-almost-40percent-of-global-employment-worsen-inequality.html https://www.cnn.com/2024/01/15/tech/imf-global-employment-risk-ai-intl-hnk/