UK Officials Said AI Would Add £400 Billion to the Economy. That Figure Mainly Came From Google

WIRED reports that the UK government's go-to figure for AI's economic upside, £400 billion, was drawn from a 2023 report commissioned by Google and produced by the consultancy Public First, now PF Global. The AI minister, Kanishka Narayan, has now told Parliament the government's own assessment puts the gain at £55 billion to £140 billion by 2030, with a central estimate of £120 billion, and called even that hugely uncertain.

The number was used across two governments to reassure people worried about job losses and an economic shock. Ministers had the lower estimate since at least November 2025, yet the larger figure kept circulating, including on a government agency's materials months later. The answer to Parliament did not withdraw the £400 billion figure or call it wrong.

This is a story about evidence, not about AI being worthless. Even the government's central estimate of £120 billion is a large number. We think AI can deliver real gains in public services and productivity, and the people quoted here, including the committee chair Dame Chi Onwurah, say as much. The problem is what happens when a government leans on an industry-funded figure to make its case and then doesn't say so.

The damage is to trust. When an official number turns out to be more than three times the government's own projection, and its origin wasn't disclosed for almost two years, every future AI claim from that government gets discounted. Onwurah's word for it is hype, and she's right that overstating the benefits undermines the case for change.

It's fair to note PF Global's side. Its head of economics says the older model assumed faster adoption than has materialized, and that its central estimate for 2030 is now much closer to the government's. The firm says its work was independent. That may well be true. But the Ada Lovelace Institute's point stands: figures that look precise often carry assumptions and incentives, and governments should ask where a number came from, what it assumes, and who benefits before building policy on it.

Where this leads if it becomes normal is policy set by whoever pays for the most quotable study. That's a bad deal for taxpayers and, ultimately, for the AI industry too, because the backlash to oversold promises lands on the technology itself. We'd rather see governments fund independent research and then measure whether the promised gains actually show up.

Lord Holmes of Richmond asked ministers about the origins of the £400 billion claim in January 2025 and did not get an answer, according to WIRED. The Cabinet Office and Google were approached for comment.