Research reveals AI is ‘not the main driver’ of US job slowdown

Cisco announced layoffs this week. Meta cut thousands earlier this year. Google, Amazon, and countless tech firms have all blamed artificial intelligence for workforce reductions. The narrative is now dominant: AI is killing jobs.

However, there is new research from the New York Federal Reserve that challenges this narrative. In their report, they conclude that AI is "not the main culprit" of the US labour slowdown.

Researchers from the New York Fed analysed vacancies in jobs that are highly exposed to AI, including programmers, customer service representatives, and data entry keyers. They used a metric created by Anthropic economists to study hiring trends before and after the release of ChatGPT in November 2022.

If AI technology had disrupted the labour market, the expectation would have been that hiring would take off in distinctly divergent paths between AI exposure and non-exposure starting from this period.

What the researchers discovered about this was the divergence between high- and low-exposure occupations began prior to 2022. Employment listings for occupations vulnerable to AI technology were dropping even before ChatGPT was launched.

The overall trend for hiring rates, however, was positive in early 2024, with rates improving to the highest point in two years. Meanwhile, the rates of layoffs have stayed relatively low at 0.9% to 1.2%, despite companies using artificial intelligence as a reason for downsizing since 2021.

According to Goldman Sachs economist Elsie Peng, while job postings in AI-affected positions have dropped below pre-pandemic levels, labour market mismatches are declining.