ICML paper: AI firms' social impact disclosures are sparse and declining
evijit · x · 2026-09-14
An ICML paper, Who Evaluates AI's Social Impacts?, presents the first comprehensive analysis of social impact evaluation reporting, examining 186 first-party release reports (model/system cards) and 248 third-party evaluation sources, supplemented by developer interviews. It covers bias, fairness, privacy, environmental costs, and labor impacts of foundation models.
The study finds a stark division of labor: first-party reporting is sparse, often superficial, and declining in areas like environmental impact and bias, while third-party evaluators deliver broader, more rigorous coverage of bias, harmful content, and performance disparities. Since ChatGPT's mainstream commercialization, many evals have quietly disappeared from model cards.
Related event: ICML Paper Finds AI Social Impact Evaluations Shrinking(3 posts)→
More from Safety
- Bing separates AI training from AI search grounding with new site controls — OnlineInference · 2026-10-31
- OpenAI reportedly to train on users' Daybreak sessions with no opt-out — dyn___ · 2026-10-31
- Forethought: US has 9x China's AI compute, can afford to pace the frontier — willmacaskill · 2026-10-31
- unpromptedconf takeaways: security programs face forced change as reachable vulnerabilities can't be risk-accepted — scottleibrand · 2026-10-31
- Reachable vulnerabilities can't be risk-accepted away, unpromptedconf speaker argues — scottleibrand · 2026-10-31
- AI safety researchers argue labs' silence on x-risk is a large-scale ethical mistake — DavidSKrueger · 2026-10-31