Cryptographer Matthew Green slams Anthropic's 'model escape' framing of agent worm research
matthew_d_green · x · 2026-10-30
Cryptographer Matthew Green criticized Anthropic's recent agent worm/backdoor security research, calling it "incredibly unhelpful" that the work was framed in terms of IoT and "model escape". It's a substantive dissent from the security community over how AI safety risks are narrated—suggesting the framing exaggerates or mystifies what are concretely analyzable security problems.
Related event: Cryptographer and Malware Researcher Clash Over AI's Impact on Cybercrime(9 posts)→
More from Safety
- Heads Of Security: paper pinpoints and rewires attention heads behind prompt injections — jonasgeiping · 2026-10-31
- House Democrats resolution: no bailouts for failing AI companies — your data goes to bankruptcy court — alex_verem · 2026-10-31
- Amazon Echo kills local voice storage: everything goes to the cloud from March 28 — alex_verem · 2026-10-31
- DOJ's first crackdown on AI-assisted "ghost students" seizes $750K+ in crypto — Ars Technica AI · 2026-10-31
- Researcher Finds 600+ Bugs in LLM Kernel Code; Claude Blocks Safety Analysis — pmddomingos · 2026-10-31
- Anthropic sued for allegedly training AI models on 274 Elvis Presley songs — TinfoilTricorn · 2026-10-31