UK AI minister urges contingency planning for possible job losses

UK AI minister Kanishka Narayan says the government should prepare for potentially severe job disruption from artificial intelligence, even though there is no evidence so far of an overall fall in employment caused by the technology. Speaking at a Labour Party conference fringe meeting in Liverpool, he described planning for a worst-case outcome as one of his priorities. The distinction matters for workers and employers: the warning concerns a possible future shock, not a finding that mass job losses have already happened.

The government must take seriously the possibility of an ‘unprecedented impact on the jobs market’, Narayan said, according to BBC News. ‘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,’ he told the meeting. He raised potential changes to labour-market policy and aspects of regulation, but the BBC account did not identify an adopted package of measures or a published contingency plan.

What the minister has proposed

Narayan said workers should have a voice as employers adopt AI. He promised ‘much better transparency’ around the introduction of AI and other technology, so that productivity gains could be measured. That would help establish what is changing inside workplaces, a question that broad employment figures alone may not settle. His remarks set out priorities for planning and oversight; they did not specify when a plan would be completed or which policy changes the government would choose.

The minister also pointed to a possible gap between company surveys and what workers are already doing. He described ‘silent adoption’ of AI by employees that might not be reflected in industry responses. While firms had reported no aggregate effect on the labour market, one technology company told him it had hired fewer people because tasks had been automated. Narayan said the net impact might therefore not be fully captured. A report of reduced hiring at one firm does not establish an economy-wide fall in jobs.

What the job-loss figures measure

The eight-million figure has an important history. In an analysis published in March 2024, IPPR examined 22,000 tasks across the UK economy and modeled several possible outcomes from deeper AI adoption. Its full-displacement worst case produced 7.9 million job losses and no gain in gross domestic product. A central scenario produced 4.4 million losses alongside a 6.3% GDP gain. A full-augmentation scenario, in which workers adapted with AI rather than being replaced, produced no job losses and a 13% GDP gain.

Those outcomes were illustrative scenarios, dependent on how AI was used and on policy choices; they were not counts of jobs already lost or a prediction that the worst case must occur. IPPR also modeled a separate, less disruptive set of outcomes for existing generative AI. Jung acknowledged at the conference that the think tank’s stark 2024 scenarios were ‘to some extent’ wrong, while arguing that AI had advanced from routine tasks towards more complex reasoning and creative work.

The distinction between tasks and jobs matters too. IPPR’s 2024 study estimated that 11% of UK work tasks were exposed to the first wave of generative AI, rising to 59% if businesses integrated the technology more deeply. At the conference, Jung gave newer estimates: about 11% of jobs were highly exposed to replacement by chatbots, while more sophisticated AI agents could affect about 60% of tasks. The newer figures were reported as Jung’s estimates; the sources do not establish corresponding observed job losses.

Who could face the greatest change

IPPR’s 2024 analysis identified back-office, entry-level and part-time work as particularly exposed in the first wave, including secretarial, customer-service and administrative roles. It said women were more likely to hold affected jobs and that young workers could face pressure if employers filled fewer entry-level positions. At the conference, Jung said women were 40% more exposed to AI-related job threats than men because of their greater representation in service work. That newer percentage is his attributed estimate, not a measured loss of employment.

Jung called for support to retrain service-sector workers and suggested measures that could slow a disruptive transition. One example was changing tax structures to make retaining staff more attractive than automating roles. He cautioned that policies for a severe disruption could be expensive. IPPR’s earlier proposals included a job-centred industrial strategy, help with job transitions, incentives that favour using AI to assist workers, and rules preserving human responsibility in areas such as health. These are proposals from the think tank, not measures the minister announced as government policy.

Prospect trade union general secretary Mike Clancy welcomed the commitment to give workers more say, according to the BBC, but warned against creating conditions in which working people become ‘victims’ of AI. The response underlines a practical question for any future plan: how workers will be involved when employers decide which tasks to automate and how the resulting gains will be shared.

What remains to be decided

Narayan’s remarks establish that he wants the government ready for a severe employment shock. They do not establish that such a shock is underway, that a contingency plan has been published, or that the government has adopted the tax, training or regulatory ideas discussed at the meeting. For now, the evidence presented by the minister and IPPR supports preparedness for several possible outcomes, while the scale and timing of any actual job displacement remain uncertain.