The UK government has admitted the £400 billion economic boost it promised from artificial intelligence was based on a report paid for by Google. The figure was not the result of official state analysis. The AI minister, Kanishka Narayan, confirmed last week that the government’s own internal estimate sits at £120 billion. This is less than a third of the widely cited number.
Narayan told parliament the central figure could range between £55 billion and £140 billion by 2030. He described this range as hugely uncertain. The £400 billion claim came from a 2023 report commissioned by Google and produced by the consultancy firm PF Global.
Jonathan Dupont, head of economics at PF Global, explained the gap between the two numbers. He argued the government was moving slower than the technology. Dupont said their 2023 model assumed a faster rate of adoption than the government currently uses. He noted that as understanding of the field grows, the assumptions behind economic models must be updated.
Government departments used the Google figure for nearly two years under two different administrations. Ministers had access to the lower estimate since at least November 2025. Skills England, an agency responsible for apprenticeships, continued to publish the £400 billion figure four months after that date. Narayan’s answer is the first public record of the government placing the two numbers side by side.
The minister did not withdraw the £400 billion claim or state it was wrong. Dame Chi Onwurah, chair of the Commons Science, Innovation and Technology Committee, told WIRED the government is right to promote AI for public services. She said the reports highlighted hype as a central barrier to digital transformation. Onwurah argued that overstating benefits risks undermining the case for change. She added that the government must be realistic and transparent about potential savings.
PF Global stated that while their view on overall potential has not changed, the slower rate of adoption in practice lowers the central estimate. They expect growth to ramp up in the years after 2030. Lord Holmes of Richmond, the Conservative peer who asked ministers about the origins of the claim in January 2025, said it was unfortunate the correction came so late. He noted that billions of pounds of public money depend on robust data.
The dispute highlights the line between government support and boosterism. Former technology minister Peter Kyle used the £400 billion figure while promoting the technology. Many parliamentarians outside the government raised questions about potential issues. Sumedha Deshmukh of the Ada Lovelace Institute said it is concerning that ministers cited a figure from Google-funded research for almost two years. She noted that precise-looking figures often have assumptions and incentives built into them.
Deshmukh said the government should test numbers before building policy on them. She called for independent research and simple questions about where figures come from and who benefits. The Cabinet Office and Google were approached for comment. PF Global said their work for Google continued but all research was independent and all judgments were their own.
What it means
For people making things or running businesses, the specific number matters less than the message. The government is no longer promising a massive, guaranteed windfall from artificial intelligence. The official stance has shifted to a more cautious view of economic impact. This means public spending on AI projects may face tighter scrutiny. Expect fewer grand announcements about guaranteed returns on investment.
