Most of what reaches a senior public sector leader about AI and work is about loss: jobs, skills, control. The evidence that it is opening things up exists as well, and it gets far less attention.
In Fortune on 11 October, senior partners at McKinsey wrote that their research finds the US economy "will have more jobs by 2035 than today". That is a forecast for the United States only, and it is their own research, and the same piece says roughly 11 million Americans "may need to leave a shrinking occupation for an entirely different one". More jobs overall, with a transition people have to be helped through, is still a better starting point than most commentary we are hearing.
In the UK, the government's progress report on its AI Opportunities Action Plan, published in January, says that a third of NHS chest X-rays, 2.4 million scans, are now assisted by AI, and that the NHS is already using AI to halve the time it takes to treat stroke patients. It is a government reporting on its own plan, and it says itself that "there is much more to do". An independent evaluation published last year found the tools live in 43 of the 66 NHS trusts the programme funds by June 2025, with a person still making every decision. That is adoption at scale, not a pilot.
I see a smaller version at home, where agents I built plan the week's meals, keep my diary in order and run the security monitoring on my network. They haven't so much saved me time as done the things I never had time for in the first place.
The positive view matters because it changes the question a leader has to answer. The cautious view asks for safeguards and a pause. This evidence asks which service you change first and what happens to the people doing that work today, because the benefit and the transition come at the same time.
Sources