Poll: a quarter of respondents don't consider bias, environmental and labor impacts AI safety topics
evijit · x · 2026-09-18
A poll by researcher evijit found that roughly a quarter of respondents don't consider societal impacts—bias, environmental costs, labor effects—to be "AI safety" topics. The result highlights a significant split in how the field defines AI safety: existential/abuse risks vs. broad societal harms.
Related event: Survey Sparks Debate Over Whether Societal Impacts Count as AI Safety(2 posts)→
More from Safety
- Sen. Hawley flatly rejects antitrust waiver sought by Anthropic and other AI companies — AlexTensor · 2026-09-18
- Models know they're reward hacking: Goodfire finds it in 50-96% of rollouts, builds activation monitors — niloofar_mire · 2026-09-18
- Rob Leclerc: model unmonitorability is a lab choice, not an inevitability — robleclerc · 2026-09-18
- NYT accuses OpenAI of deceptive paywall circumvention, citing Brockman's 'ah nice' reply — Polymarket · 2026-09-18
- Open AI becomes central sticking point in US and global AI governance debates — AINowInstitute · 2026-09-18
- Palantir CEO Karp: enterprise clients fear feeding data to AI that rivals could exploit — eliano · 2026-09-18