Danish study challenges claims about AI disrupting the labor market

New research suggests generative AI tools like ChatGPT have yet to make a meaningful impact on employment or wages despite their rapid adoption. A comprehensive study of the Danish labor market in 2023-2024 found no significant economic effects across occupations considered vulnerable to automation, challenging popular narratives about AI’s immediate transformative potential in the workplace.

The big picture: Economists from the University of Chicago and University of Copenhagen analyzed data from 25,000 workers across 11 occupations theoretically vulnerable to AI disruption but found essentially zero impact on earnings or work hours.

• The researchers’ statistical analysis ruled out average effects larger than 1% on wages or employment during the study period.
• The findings contrast starkly with both apocalyptic predictions of job losses and optimistic forecasts of productivity booms from early generative AI adoption.

Key details: The study, titled “Large Language Models, Small Labor Market Effects,” examined accountants, software developers, customer support specialists, and other occupations where AI tools have seen significant uptake.

• Employer-encouraged adoption was widespread across the studied fields, with most workers in vulnerable occupations now using AI chatbots.
• Corporate investment successfully boosted AI tool adoption among employees, with 64-90% of users reporting time savings.

Behind the numbers: Despite widespread adoption, productivity gains were modest, with users reporting average time savings of just 2.8% of work hours—approximately one hour per week.

• Only 3-7% of these modest productivity improvements translated into higher earnings for workers, raising questions about who primarily benefits from AI efficiency gains.
• The findings contradict a February randomized controlled trial that found generative AI increased worker productivity by 15% on average.

Counterpoints: AI tools aren’t simply saving time—they’re creating new job tasks for 8.4% of workers, including those who don’t directly use the technology.

• Teachers now spend time detecting ChatGPT use in student work, while other professionals dedicate hours to reviewing AI output quality or developing effective prompts.
• The study’s co-author suggests laboratory experiments often focus on idealized tasks highly suited to AI, whereas real-world jobs involve many functions AI cannot yet automate.

Where we go from here: The Danish study provides a valuable early snapshot but has important limitations in assessing AI’s long-term impact on labor markets.

• Data from 2023-2024 captures only the initial phase of generative AI deployment, potentially missing lagging effects or more sophisticated implementation approaches beyond basic chatbot use.
• The Danish labor market’s specific characteristics may not reflect impacts already materializing in other regions or specialized fields like freelance creative work.