The £400 billion that British ministers have said AI will add to the economy came from a report paid for by Google, and the government’s own estimate is far lower. WIRED reported the gap on 4 November, after AI minister Kanishka Narayan answered a parliamentary question about where the number began.
Narayan said the figure was taken from a 2023 report that Google commissioned and Public First wrote. Public First is a British consultancy now called PF Global. The government’s more recent assessment puts the gain from AI adoption and data investment at between £55 billion and £140 billion by 2030, with a central estimate of £120 billion. Narayan described even that as “hugely uncertain”.
The big number did a lot of work. WIRED says it was meant to calm fears that AI would cost jobs or even trigger an economic crash, and government bodies kept citing it across two administrations, for close to two years. Former technology minister Peter Kyle cited it while promoting the technology. WIRED says the lower number reached ministers by November 2025 at the latest. Even so, Skills England, the agency that handles apprenticeships, kept the old number on its materials four months later.
WIRED found no earlier record of the government placing the two estimates together or naming Google as the client. Narayan’s answer also left the £400 billion standing, neither retracting it nor calling it mistaken.
Jonathan Dupont, who runs economics at PF Global, told WIRED the difference comes mostly from adoption speed. The government’s numbers assume a much slower take-up than looked possible when the 2023 model was built. “AI is a very fast moving technology,” he said, so assumptions have to be updated. PF Global says its own central estimate for what gets achieved by 2030 is now much closer to the government’s, with growth picking up in the years after. It says its view of the technology’s overall potential has not changed, and that its work for Google continues, with all judgments and assumptions its own.
Politicians and researchers quoted by WIRED were less forgiving about the delay. Dame Chi Onwurah, a Labour MP, chairs the Commons committee on science, innovation and technology. She said overstating benefits “risks undermining the case for change”. Lord Holmes of Richmond, a Conservative peer, said he asked ministers about the origin of the £400 billion in January 2025 and got no response. He told WIRED it would have been better for the government to clarify “much sooner and more directly”.
Sumedha Deshmukh, from the research body Ada Lovelace Institute, called the revision welcome but pointed to a pattern. Precise-looking AI figures often carry built-in assumptions and incentives, she said, and industry often funds or shapes the research behind them. Her advice to government: ask where a number came from, what it assumes, and who benefits from it before building policy on it. WIRED asked the Cabinet Office and Google for comment, and its account does not include a reply from either.
The episode shows how a sponsored forecast can harden into official fact. A consultancy’s 2023 model, built for a company with an obvious interest, became a talking point across two governments before anyone published the alternative. The revised range is itself wide, an £85 billion spread between low and high, which is hardly more reassuring than a single confident headline.
Anyone who has copied the £400 billion into a pitch deck, a funding bid, or a lobbying brief now has a government figure of £120 billion to replace it, and a Commons committee chair on record warning about hype.
Reported by WIRED (Ben Gartside) on 4 November 2026.
