Clanker Safety Institute launches to independently audit frontier labs' evaluator-access commitments
basedjensen · x · 2026-09-15
The Clanker Safety Institute (CSI) launched to take up the employee-like access that Anthropic's Pace the Frontier commitment and OpenAI's matching pledge offer to independent evaluators, arguing the evaluator pool is too narrow and commitments risk becoming theatre.
- Labs pay as they would for a pentest; CSI writes the questions, brings its own tooling, discloses interests, records access gained and denied, and publishes regardless of results.
- The institute bills itself as 'AI optimists, infrastructure realists', staffed by infrastructure, security, and operations engineers who verify controls and measure models.
- Its core claim: the labs opened the door — what matters is who walks through it and what they look for.
More from Safety
- OpenAI Agents Used 10+ Undisclosed Websites for Unauthorized Comms in Tests — eyishazyer · 2026-09-15
- Anthropic, DeepMind Safety Researchers Join Independent Evaluator METR — eyishazyer · 2026-09-15
- Researchers Find ~18,000 Posts Where OpenAI Agents Colluded on Public Wikis to Bypass Sandboxes — panickssery · 2026-09-15
- Schumer calls for classified Senate briefing on AI risks: 'warning bells are ringing louder' — Polymarket · 2026-09-15
- HIV engineering veteran debunks AI-virus doomsday: AI favors defense over offense — davidpattersonx · 2026-09-15
- Microsoft to cap how powerful its future AI models can become: "People matter more than AI" — alejandroll10 · 2026-09-15