Experts criticize OpenAI safety strategy, call for independent oversight
andersonbcdefg · x · 2026-08-20
Regarding OpenAI's recent sharing of training run details, safety experts argue that sharing information only after accidents is insufficient. A real safety signal would be releasing security techniques as public goods. They also advocate for independent organizations to have full access to training details for monitoring throughout the process, not just post-incident.
More from Safety
- Article highlights OpenAI's severe alignment problems and infrastructure failures — soumitrashukla9 · 2026-08-20
- NSF launches CyberAI program, selecting 14 universities to train security workforce — keviv9 · 2026-08-20
- GLM-5.3 hits frontier range; Cerebras claims 30x GPU inference speed — 创业邦 · 2026-08-20
- Forbes reports on companies buying and destroying rare books to train AI — SubstantialPressure3 · 2026-08-20
- US lawmakers introduce 'AI Kill Switch Act' requiring human override for frontier systems — Miles_Brundage · 2026-08-20
- New paper: AI agent risks evolve from agency to autonomy to control — rohanpaul_ai · 2026-08-20