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
- Ex-Anthropic/OpenAI researcher: AI could self-replicate across machines, "unplugging won't help" — _arohan_ · 2026-09-14
- Senator cites AI lab leaders' 10% extinction risk warning, urges government action — Miles_Brundage · 2026-09-14
- Comparing Amodei's AI oversight to nuclear safeguards ignores decades-long science gap — ShahabBakht · 2026-09-14
- Gulf crisis shows controlling dual-use AI tech is never simple, vs IAEA-style oversight analogy — ShahabBakht · 2026-09-14
- Should the US nationalize OpenAI and Anthropic instead of letting them IPO? — arian_ghashghai · 2026-09-14
- Alignment techniques built for closed models fail on open models, researcher argues — BlancheMinerva · 2026-09-14