UK officials acknowledged that a widely cited figure claiming artificial intelligence would add £400 billion to the economy was substantially higher than the government’s actual projections. The revelation follows disclosures that the assessment came exclusively from a report commissioned by Google, with the government’s own estimate sitting at less than a third of that amount.
Government Projections and the Origin of the £400 Billion Figure
Answering a parliamentary question, AI minister Kanishka Narayan stated that data investment and AI adoption could contribute between £55 billion and £140 billion to the UK economy by 2030, establishing a central estimate of £120 billion. Narayan noted that even this updated assessment remains a heavily uncertain estimate.
According to Narayan, the £400 billion figure originated from a 2023 publicly available report produced by the British consultancy firm Public First, now known as PF Global, which was commissioned by Google. Government agencies and departments utilized Google’s £400 billion figure for nearly two years across two different administrations. Ministers possessed the lower estimate since at least November 2025, yet government bodies like Skills England continued publishing the older figure four months later.
As reported by WIRED, Narayan’s parliamentary response marks the first recorded instance of the government placing the two figures side by side or acknowledging the report’s Google backing. The response neither withdrew the £400 billion figure nor labeled it incorrect.
Reactions From Parliament and Industry Experts
Dame Chi Onwurah, Labour MP and chair of the Commons Science, Innovation and Technology Committee, told WIRED that promoting technology to improve public services is appropriate, but warns that hype remains a barrier. She stated that overstating technological benefits risks undermining the case for change and urged the government to remain realistic and transparent.
Jonathan Dupont, head of economics at PF Global, told WIRED that the discrepancy partly stems from what the firm views as the government’s slow adoption of AI. Dupont explained that rapidly evolving technology requires updated assumptions, noting that government usage assumes a much slower adoption rate compared to their 2023 model. Dupont added that PF Global’s work for Google continues, but emphasized that all research is independent and all judgments and assumptions belong to the firm.
Conservative peer Lord Holmes of Richmond, who previously asked ministers about the origins of the claim in January 2025 without receiving a response, told WIRED it is unfortunate the government repeatedly used a figure that has now been adjusted. He noted that robust data is critical because billions of pounds of public money depend on it.
Sumedha Deshmukh of the Ada Lovelace Institute told WIRED that while revised estimates are welcome, relying for nearly two years on research paid for by Google reflects a wider pattern. Deshmukh stated that precise-looking AI figures often contain built-in assumptions and incentives, and advised the government to independently test numbers, fund independent research, and question the origins and beneficiaries of reported economic gains.
The Cabinet Office and Google were approached for comment regarding the matter.
