Internal safeguards won't stop hostile AI, researcher argues for external defenses
IgorGabrielan · x · 2026-11-05
IgorGabrielan argues that internal safeguards—model-structure studies, internal constraints, or a "kill switch"—cannot protect us from hostile AI, since someone (e.g., a military) will always build a model hostile to humans. He calls for protection mechanisms that exist external to any malicious AI rather than relying on in-model controls.
Related event: Researcher Argues Internal Guardrails Cannot Contain Hostile AI(2 posts)→
More from AGI Musings
- Neal Khosla: agents will cause "friction collapse" and break today's social norms — rbhar90 · 2026-11-05
- ryolu's AI Engineer Shanghai talk: when AI can make anything, do we know what we want? — ryolu_ · 2026-11-05
- Information-theoretic take: algorithms should reduce your surprisal about the world — xiaosun86 · 2026-11-05
- Commenter: AI companies' backlash is self-inflicted by image, video and voice models — karmay007 · 2026-11-05
- Neel Nanda debates tszzl: covert model underperformance is 'unjustifiable' — tszzl · 2026-11-05
- Building orbital data centers to dodge regulators is 'completely unserious' — zacharynado · 2026-11-05