AI is not yet replacing workers in the US, researchers find (www.cnn.com)

A new study from Yale’s Budget Lab finds that ChatGPT and the broader wave of generative AI have not yet produced a measurable disruption in the U.S. labor market. By tracking changes in the distribution of workers across all occupations since ChatGPT’s public release (33 months), researchers tested whether generative AI has pushed people between jobs, automated them out of roles, or created new categories of work—and found no discernible, economy‑wide shift to date. The paper stresses this is an early‑stage snapshot rather than a prediction, and the team will continue monthly monitoring to detect any emerging trends. The finding tempers high‑profile warnings of mass unemployment from some industry leaders, but it coexists with mixed signals: individual firms (e.g., Dropbox, Duolingo) have cited AI as a factor in layoffs, surveys show many employers plan AI‑driven downsizing, and new work patterns like “workslop” (AI generating passable but error‑prone outputs that require human correction) are surfacing. For the AI/ML community this underscores two technical implications: adoption and impact are uneven and task‑specific, and real productivity effects require rigorous measurement and workflow evaluation. Ongoing, granular monitoring and better tools for human–AI collaboration, verification, and auditing will be critical to understand whether and how generative models reshape employment over time.