Reporting archive
Evidence on AI and work
-
OpenAI’s firing of three safety researchers sets off a dispute over trust, leaks, and AI safety work
OpenAI said it fired three safety researchers, while the researchers said the move chilled open safety work and asked for stronger outside oversight. The dispute centers on sensitive information, delegated email access, and what AI safety teams can share.
-
New Census paper links AI exposure to weaker first-job outcomes for college graduates
A Census working paper finds the sharpest AI-linked labor-market effects at college graduates’ first job, with weaker initial employment and earnings in the most exposed majors.
-
Meta AI-layoff fight returns to court in California
A federal judge is hearing a request for preliminary relief in a case that tests whether AI-assisted workforce tools can be challenged in layoff selection. Meta says people, not AI, made the decisions.
-
Bank of Canada says AI is starting to make hiring harder in exposed occupations
The Bank of Canada says AI is making it harder for some Canadians to find work in occupations most exposed to it, even as it does not see broad changes in the overall labor market structure.
-
No fresh AI-and-jobs event could be verified in the last 48 hours
No qualifying AI-and-jobs development could be verified from the last 48 hours. The closest evidence-backed item was Revelio Labs’ AI Labor Market Tracker launch on July 28, 2026, which falls outside the requested window.
-
Apollo’s AI paper points to weaker pay growth, not just layoffs
Apollo Global Management’s new white paper, as reported by Axios, argues that AI exposure is showing up in slower wage growth for some occupations more than in headline job losses.
-
Meta AI-layoff lawsuit survives initial bid to block July 22 cuts
A U.S. judge declined emergency relief in Meta’s AI-layoff lawsuit, leaving the July 22 layoff timeline in place.