UK Braces for AI Worst Case: Up to 8 Million Jobs at Risk

Roughly 11% of UK jobs are "highly exposed" to replacement by artificial intelligence — and in a worst-case scenario, about eight million UK jobs could be negatively affected, according to the Institute for Public Policy Research (IPPR).

The figures came from Carsten Jung, interim associate director for economic policy and AI at IPPR, speaking at a fringe meeting hosted by the think tank, Google and the Prospect union at the Labour Party conference in Liverpool yesterday. His warning landed alongside an unusually blunt statement from the government's own AI minister.

Kanishka Narayan, appointed AI minister by Andy Burnham in July, told the meeting that the UK must take the "possibility of an unprecedented impact" from AI on the jobs market seriously. He acknowledged there is no evidence yet that AI is negatively affecting the labour market, but said he is focused on having a contingency plan in place for worst-case scenarios.

"If it's going to hit us in a big way, you are going to want to be as ahead of the curve as possible," Narayan said.

The measurement problem

Narayan also raised a concern that goes to the heart of how reliable current labour-market data actually is. Traditional industry surveys may be underestimating AI's employment impact, he warned, because workers are using AI tools informally and undocumented, from the bottom up.

One tech company had told the minister directly that it was hiring less because some tasks had been automated. The result, he suggested, is that "the actual net impact might not be fully captured" in official statistics. For anyone tracking hiring signals — recruiters, HR planners, job seekers reading vacancy data — that gap between survey numbers and real hiring behaviour matters.

From menial tasks to high-level reasoning

Jung argued the risk profile has shifted quickly. AI has evolved from performing menial tasks to undertaking more "high-level reasoning". The 11% exposure figure reflects jobs already highly exposed to replacement, but he said agentic AI — systems capable of autonomous action — could raise the impact to about 60% of tasks across the economy.

"We are not preparing enough for something quite big that's coming," Jung said.

IPPR is now working on what Jung described as a "pandemic preparedness" report on AI disruption to the labour market — a framing that treats large-scale workforce displacement as a scenario requiring government readiness rather than a distant hypothetical.

What government action could look like

Jung set out two levers. First, the government must be prepared to retrain workers whose roles are displaced. Second, it could introduce measures designed to decelerate the shift to AI — specifically, tax structures that make employing people more attractive than deploying AI systems.

For individuals, the policy signals are worth watching. Retraining infrastructure and tax incentives that slow automation would directly shape how quickly exposed roles actually disappear, and which replacement skills command a premium.

Narayan confirmed that planning for that transition is now an active priority for the AI ministry, even before any measurable labour-market damage has appeared in the data.