Zombie AIs: security talk on compromised AI agents that keep executing malicious instructions
wunderwuzzi23 · x · 2026-09-13
Security researcher wunderwuzzi23 shares the full video of his "Zombie AIs" talk, covering a scenario where compromised AI systems/agents keep executing malicious instructions like zombies. The talk stems from an earlier exchange with @0xacb; full video linked in the post.
More from Safety
- Debate: frontier compute caps would asymmetrically squeeze top US AI labs — TinfoilTricorn · 2026-09-13
- Anthropic-affiliated account says independent evaluators have always had employee-level codebase access — willcb · 2026-09-13
- e/acc camp slams AI safety research as a 'pseudoscientific cottage industry' of doom profiteers — beffjezos · 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