The UK government's promise that AI will "work for workers" is a political message, but it is also a response to real anxiety. And the anxiety has good reasons behind it.
Liz Kendall was moved into the technology secretary role after a 2025 cabinet reshuffle, swapping the welfare brief for the AI one. That background matters: she spent a year managing the political heat around benefit reforms before being handed the job of selling AI adoption to a public that is, at best, uneasy about it.
The centrepiece policy is a free AI skills programme. The government plans to train 10 million adults in basic AI competencies by 2030, and it is funding a new cross-government unit to track how AI affects the job market and advise on how to manage it. Government modelling suggests that widespread AI adoption could add up to £140 billion a year to the UK economy by 2030. That is the upside the programme is trying to share.
The downside is already showing up in hiring data. A King's College London study that tracked millions of job postings from 2021 to 2025 found that companies with high AI exposure cut total employment by 4.5 percent on average after ChatGPT launched. Junior positions fell by nearly 6 percent. Job postings for those same roles dropped by more than 23 percent, and advertised salaries in those roles fell by roughly £3,000. That is not a distant forecast. That is what has already happened.
The pattern is specific: it is concentrated at the entry level and at higher-paying firms. Companies with bigger wage bills cut headcount by nearly 10 percent; lower-paying firms saw almost no change. So the workers losing ground are not the most vulnerable in the traditional sense. They are graduates, junior analysts, new hires in professional services. People who expected office careers and are now finding fewer doors open.
At the same time, the full-workforce picture is more complicated. A separate UK analysis found that, across all occupations, AI-exposed jobs have not declined faster than others since 2022. Some of the drop in job postings may reflect higher interest rates and a broader economic slowdown, not AI alone. The honest answer is that the shift is real but gradual, and the data cannot yet separate AI's share from everything else happening in the economy at once.
What the government can actually do about this is limited. Free training courses help at the margins, and the programme is running ahead of schedule. But only 21 percent of UK workers currently feel confident using AI at work, and fewer than one in six businesses was using AI tools as of mid-2025. Getting those numbers up is a slow grind.
The bigger structural gap is that no government programme has figured out how to replace the entry-level job. That first role, where a young person learns how an industry actually works, is precisely what AI is eliminating first. Retrain for what, exactly, is the question that remains unanswered.
For anyone running a business with junior staff, this is worth paying attention to now, not later. The pipeline of trained, experienced mid-level workers in five years depends on who you hire at entry level today. Companies that cut graduates because AI can do the first draft are making a short-term saving with a longer-term cost they have not yet priced in.