SerpApi.Org
AI does not appear to be boosting productivity, analysts at Barclays have said.
Evidence that adopting the technology makes workers more productive was “unconvincing”, the bank’s experts said in a note sent to clients on Tuesday.
Barclays found little to suggest that US industries that had embraced AI had seen larger improvements to productivity growth than before adopting the technology.
“Industry-level evidence linking AI adoption to stronger productivity remains weak,” analysts wrote.
The finding will fuel concerns about a potential stock market bubble inflated by AI. US markets crashed last week on concerns that the trillions of dollars being spent on AI infrastructure would not pay off.
Companies around the world have enthusiastically adopted models such as OpenAI’s ChatGPT, Google’s Gemini and Anthropic’s Claude in the hope of spurring their businesses.
Recommended
China's baby-faced billionaire sends global markets into panic
However, some have sought to temper their enthusiasm in recent months.
Uber moved to reduce its spending on AI earlier this year, limiting all employees to $1,500 (£1,100) in monthly token spending per AI coding tool. It came after the company blew through its AI budget earlier this year.
The boss of Klarna admitted last year that his pursuit of cost-cutting in customer service, fuelled by advancements in AI, had gone too far. The fintech stopped hiring for more than a year as it focused on building AI capabilities, but it started rehiring people in 2025 for its customer service operations.
Mark Zuckerberg, the chief executive of Meta, also recently admitted that the social media giant had made mistakes in its AI-driven transformation of its workforce. The Facebook founder laid off 10pc of its workforce and transferred 7,000 employees to new initiatives in AI.
Kevin Warsh, the new chairman of the Federal Reserve, has heralded an AI productivity boom as a possible path to lowering interest rates in the US.
Mr Warsh characterised the AI boom as “the most productivity-enhancing wave of our lifetimes – past, present and future”.
He said earlier this year that AI would be “structurally disinflationary”. He has argued that AI-driven productivity would lower the cost of goods and do away with the need for higher interest rates.
Barclays analysts said this had so far not materialised: “Our bottom line is that AI adoption appears gradual and steady rather than rapid and transformative, with most households and businesses still reporting limited exposure to the technology.
“At the same time, evidence of a structural pickup in productivity growth remains surprisingly fragile.”
Other economists believe that, at least in the short term, AI could be inflationary. Memory chip prices have soared because of the strong demand in data centres, pushing up the costs of smartphones and other gadgets.
David Kelly, the chief global strategist at JP Morgan, said in April: “It does appear that AI is, on balance, adding slightly to inflation in the short run.”
The study by Barclays found that while some sectors such as professional services had seen widespread adoption of AI, others such as public administration and retail had not been so quick to start using the technology.
Surveys in the US suggested that roughly half of adults in the country were using AI in some capacity, Barclays said.
Most appear to still be using it sparingly. Just 14pc of respondents to a nationwide survey in the US said they were using AI daily for work, as of the second quarter of 2026.
