Agent misuse risk is set at deployment: tool access and auditability deserve scrutiny
AlexTensor · x · 2026-10-23
Responding in a debate involving Yoshua Bengio, Andres Dancausa argues that developer decisions don't stop at the weights: for agents, much of the misuse risk is set at deployment — which tools the agent can call, what it can execute, and whether anyone can audit it afterward. That layer deserves the same scrutiny as the model itself.
More from Safety
- Beff Jezos calls for a 'cyber immune system' of constantly adaptive agents — beffjezos · 2026-10-23
- Security expert: cybercriminals aren't engineer-tier, AI-native crime syndicates will take time — joshua_saxe · 2026-10-23
- Pedro Domingos predicts 2027 will see the most security bugs fixed in history — pmddomingos · 2026-10-23
- AI safety researchers argue even finance-grade governance may not be enough for AI — geoffreyirving · 2026-10-23
- MalwareTech: LLM-hallucinated php2shell exploits flooding GitHub don't actually work — paul_cal · 2026-10-23
- Security researcher bets AI won't formally verify Linux kernel, Chromium and other critical software by 2028 — JeffLadish · 2026-10-23