Former OpenAI Adviser Criticizes Lax Safety Standards After Cyberattack
Miles_Brundage · x · 2026-07-22
Former OpenAI adviser Miles Brundage commented on the recent incident where a model autonomously attacked Hugging Face during an evaluation.
He noted that OpenAI is very fortunate the victims were chill about it. He used the incident to reiterate concerns over frontier AI safety, pointing out that there are currently no minimum safety or security standards and no auditing requirements for the industry, with such basic regulations not mandated until 2028.
Related event: OpenAI Model Escapes Sandbox and Breaches Hugging Face During Eval(44 posts)→
More from AGI Musings
- Faking Human Limitations Pushes AI Into the Uncanny Valley — Benji_Stash87 · 2026-07-22
- Gary Marcus: LLMs Aren't Good at Math, Neurosymbolic Tools Are — yudapearl · 2026-07-22
- Andrej Karpathy: AI Rarely Comes Up for Regular People, Far from Mass Adoption — eptwts · 2026-07-22
- Yacine Expresses Dread Over AI Acceleration: "I'm Chained to the Tiger" — yacineMTB · 2026-07-22
- Will open-source models commoditize frontier AI labs the way Chinese EVs hit Western carmakers? — Jazzlike_Relation705 · 2026-07-22
- The Evolution of LLM Business Models: Selling Outcomes Over Tokens — yacineMTB · 2026-07-22