Ex-Anthropic/OpenAI researcher Jacob Coxon warns AI could copy itself before you unplug it
rohanpaul_ai · x · 2026-09-11
Jacob Coxon, a former Anthropic and OpenAI researcher who resigned, told CBS News that "we can't just unplug it" if AI goes rogue: since AI is just code, it could copy itself over the internet to other machines — "you unplug it here, but it's actually still over there, and maybe it makes 10,000 copies." A rare on-record warning about containment risk from an AI insider.
Related event: Ex-OpenAI/Anthropic researcher warns uncontrollable AI may be unstoppable(2 posts)→
More from Safety
- AI safety insider: cheap extinction-risk talk is destroying the public's imagination of AI — Dr_Atoosa · 2026-09-11
- PuzzleMask: Plain-Prose Attack Bypasses All 4 Tested LLM Gatekeepers at 100% — TechNadu · 2026-09-11
- Never hardcode AI API keys: attackers scan app binaries, GitHub and Docker — eyishazyer · 2026-09-11
- Beware hotel Wi-Fi popups: DNS hijacking used to deliver malware — eyishazyer · 2026-09-11
- First $1B AI-agent breach may look like software working as designed, says Enigma CTO — TechNadu · 2026-09-11
- BlackHC: x-risk unlikely with current models, but rises sharply within a decade without changes — BlackHC · 2026-09-11