Clanker Safety Institute launches to audit frontier labs' evaluator-access pledges
basedjensen · x · 2026-09-15
After Dario Amodei's "pace the frontier" pledge and Sam Altman's matching commitment to give independent evaluators employee-like access, the Clanker Safety Institute launched as an independent auditor built for that access. Staffed by infra/security/ops engineers, it audits deployment controls (isolation, default-deny egress, least privilege, change control) alongside model capabilities and safety behavior. Labs pay as for a pentest; CSI brings its own tooling, discloses interests, and publishes regardless of results — arguing verified controls buy certainty where fixed slowdowns only buy time.
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