AVERI Paper on Frontier AI Auditing Backs Dario's Embedded Evaluator Commitment
Miles_Brundage · x · 2026-09-13
AVERI, joined by experts including Yoshua Bengio and Rishi Bommasani, published "Frontier AI Auditing," a framework for rigorous third-party assessment of leading AI companies, with embedded auditing (deep, secure access to non-public information) as a key component. The group commended Anthropic and OpenAI's commitments and argues that making embedded auditing universal requires standards, tooling, and policy design beyond pilot projects.
More from Safety
- Debate: frontier compute caps would asymmetrically squeeze top US AI labs — TinfoilTricorn · 2026-09-13
- OpenAI's army of bug-hunting agents sparks AI safety firestorm: 'false flag' vs 'what a farce' — beffjezos · 2026-09-13
- Coefficient Giving launches Project Tailwind, a call for ambitious AI safety founders — EthanJPerez · 2026-09-13
- e/acc-Adjacent Lexicon: 'Safety Cartel' Defined as AI Incumbents Using Safety to Slow Rivals — PeterDiamandis · 2026-09-13
- Frontier labs fear open source + RL fine-tuning hitting near-frontier perf at 20x lower cost — beffjezos · 2026-09-13
- SALT, MAD and containment did slow nuclear proliferation — a case for the AI containment analogy — aronchick · 2026-09-13