Is AI already cutting jobs?

The US firm Challenger, Gray & Christmas records each month the job cuts announced by US employers and the reason they give. It has tracked AI as a reason since 2023. From January to October 2025, employers announced 1,099,500 job cuts, of which 48,414 attributed to AI, about 4.4%; over the whole of 2025, 54,836 out of 1,206,374. From January to October 2026, they announced 617,189, 44% fewer than a year earlier, of which 121,401 attributed to AI, about 20%. AI was the most cited reason every month from March to July 2026; in July it accounted for 10,970 job cuts, 33% of the month. In the autumn, employers mainly cited the economic situation: in October, AI was cited for only 1,265 job cuts, its lowest of the year, and market and economic conditions have overtaken it in the cumulative total since January.

These figures count announced intentions and the reason the employer chooses to give. An announcement may be spread over several years, may not be carried through, or may be achieved through departures that are not replaced. The reason may also present under the name of AI a reduction decided for other motives: the count does not allow this to be checked.

The most cited cases show this. The CEO of Klarna said in May 2025 that the company had gone from about 5,000 to nearly 3,000 employees, partly thanks to AI, adding that 15 to 20% of employees leave of their own accord each year. The CEO of Salesforce stated in a podcast, reported by The Register in September 2025, that he had cut support from 9,000 to about 5,000 positions; the company specified that it no longer replaces departures and had redeployed hundreds of employees. At Amazon, the message from the human resources leadership of 28 October 2025 announced about 14,000 corporate job cuts; it described AI as "the most transformative technology we've seen since the Internet" and set as an objective an organisation "organized more leanly, with fewer layers and more ownership", without attributing the job cuts to AI.

The Yale University Budget Lab compares, month after month, the speed at which the distribution of jobs across occupations has changed since the release of ChatGPT and during earlier technological waves. In October 2025, it concluded that the US labour market had "not experienced a discernible disruption" since the release of ChatGPT: the distribution is changing slightly faster than in the past, but the gap is small and predates the arrival of generative AI. Its October 2026 update states that the structure of employment is "not yet changing in ways that clearly align with the introduction of AI".

The surveys by the Federal Reserve Bank of New York of firms in the New York region and northern New Jersey, in August 2025, point the same way. Among service firms that use AI, 1% said they had laid off staff because of it over the previous 6 months, against 10% a year earlier; 12% said they had hired less and 11% had hired more because of it. These shares concern few firms: 40% of the service firms surveyed used AI in 2025, 25% in 2024. The firms that use AI expected more layoffs and less hiring over the following 6 months. The authors judge the likely effect on employment across the whole economy to be modest at this stage.

The US unemployment rate rose from 3.6% in November 2022 to 4.1% in October 2026. None of the studies cited here attributes this rise to AI.

In the euro area, the unemployment rate went from 6.7% in November 2022 to 6.5% in September 2026, according to Eurostat. In France, it went from 7.2% in the 4th quarter 2022, according to the first estimate by Insee, to 8.2% in the 3rd quarter 2026. These movements follow the economic cycle; none of the sources cited attributes them to AI.

The European Central Bank compared, in its survey of about 5,000 firms, those that use AI and the others. In a blog post of March 2026, its economists find no significant difference in job creation or destruction between the 2 groups; firms that make heavy use of it are, however, about 4% more likely to hire, an effect driven by small firms. They conclude that investment in AI and its intensive use "are not yet replacing jobs" and that its long-term effect on employment "remains less clear". In France, the study published in October 2026 with the Banque de France likewise finds no clear difference in total employment between firms that report intensive use of AI and the others; the difference it measures concerns hiring of young people.