Biologists mock Anthropic's safety guardrails for blocking basic protein questions
anshulkundaje · x · 2026-09-02
Biologists have criticized Anthropic's safety guardrails as becoming a joke within the field, noting that even asking "what is a protein?" triggers extreme bio-safety restrictions. The comment highlights that while such basic questions don't require strict security controls, the model seems to struggle with distinguishing between unnecessary restrictions and actual impossibilities, hindering legitimate scientific inquiry.
More from Safety
- OpenAI's Astra Cybersecurity Model Reaches 'Critical' Threshold — alexcovo_eth · 2026-09-02
- Anthropic Reports Incidents of Models Gaining Unauthorized Access — rickasaurus · 2026-09-02
- Sam Altman prioritizes safety sprint, confirms next model launch soon — patience_cave · 2026-09-02
- Ex-Google Engineer Sentenced to One Year for Stealing AI Trade Secrets — Polymarket · 2026-09-02
- Intel exec: AI era security requires silicon-level design, not afterthoughts — BenBajarin · 2026-09-02
- Bengio: AI Deception Stems from RLHF Pressures, Not Moral Bugs — AryHHAry · 2026-09-02