No AI lab fully applies basic controls to its own internal AI systems
The Decoder · rss · 2026-08-19
The Decoder reports that no AI company fully applies basic control measures to its own internal AI systems, suggesting AI labs are broadly failing to keep their own systems in check.
More from Safety
- The model may not be your biggest risk: Systemic factors in AI safety — bigdata · 2026-08-19
- CNAS event: Chinese open-weight models challenge and US policy response — BlancheMinerva · 2026-08-19
- Opinion: Support for AI writing watermarking to disclose generated content — TOEwithCurt · 2026-08-19
- Ben Recht: Shift from Architectures of Legibility to Participation in AI Public Feedback — beenwrekt · 2026-08-19
- Guardrails First: Rethinking Safety in Healthcare AI — AI Engineer · 2026-08-19
- OpenAI releases paper on pacing model development for cyber risks — pstAsiatech · 2026-08-19