Despite growing adoption, a new study reveals that AI tools, including chatbots like ChatGPT, have had “no significant impact on earnings or recorded hours” across occupations, failing to meet expectations of workplace productivity.
The study, led by Denmark researchers and covering numerous occupations, suggests that while AI may streamline certain tasks, it has not led to measurable changes in pay or work hours. These findings challenge widespread assumptions that AI will dramatically transform the labor market, indicating its impact may be more limited than many have projected.
AI saves less than 3% of time, but with no noticeable work changes
A recent study by economists Anders Humlum and Emilie Vestergaard examined the adoption and labor market effects of ChatGPT and other similar AI tools among 25,000 workers in 11 occupations in Denmark. While AI adoption is widespread — about half the workers in affected job categories using artificial intelligence — the study found no significant impact on wages or hours worked. The research estimates that AI saves just 2.8% of work time on average, or roughly one hour per 40-hour week.
Researchers also observed that workers spend additional time reviewing AI outputs or managing new responsibilities. New AI-related tasks have emerged for 8.4% of workers, further diluting the potential productivity boost. Overall, the data undercuts expectations of a rapid AI-driven transformation in labor markets.
Why AI isn’t boosting productivity as expected
Some experts argue that the underwhelming AI productivity gains, as seen in Humlum and Vestergaard’s study, are rooted in how organizations implement the technology.
MIT’s Ravin Jesuthasan contends that introducing AI alone is insufficient; real gains require companies to redesign workflows, automate routine functions, and reorganize teams. Without this fundamental change, businesses risk maintaining outdated processes, limiting the benefits of AI adoption.
Some researchers report substantial productivity gains with AI adoption
By contrast, a 2023 NBER paper by Erik Brynjolfsson, Danielle Li, and Lindsey R. Raymond revealed that generative AI tools increased productivity among customer support agents by nearly 14%, with less experienced workers seeing improvements as high as 34%.
These contrasting findings highlight the complexity of assessing AI’s economic impact which can vary widely depending on the industry, tasks structure, and workforce composition.
AI’s true labor impact
Whether AI will truly revolutionize labor remains uncertain. So far, early evidence paints a mixed picture. For now, more research — and patience — may be needed before drawing firm conclusions. Until the impact becomes clearer, purposeful AI integration may offer more than sweeping expectations.