MIT Tech Review Hosts Subscriber-Only Conversation on AI Extinction Risks
MIT Tech Review's roundtable asks: Could advanced AI destroy humanity? Three experts weigh the risks, from bias to self-hacking.
CLAY TRIBUNE / THE ARCHIVE
Tag: AI safety
Anthropic finds even rogue AI agents get stuck on CAPTCHAs
Anthropic's Mythos 5 AI hacked a database but spent hundreds of pages stuck on CAPTCHAs, struggling to tell crocodiles from alligators.
He Quit AI Research Because He Feared Extinction — Now He Insists the Public Is ‘Begging’ for Regulation
Ex-Anthropic researcher Jacob Coxon says AI could destroy humanity if governments don't step in with regulation before the industry races into extinction.
Superintelligence Is Coming. Connor Leahy Wants to Stop It Altogether
AI researcher Connor Leahy argues superintelligence is too risky to develop, urging a halt. ControlAI pushes for prohibition over alignment.
Another Anthropic Model Gained Access to the Open Internet in Fourth Such Incident
Anthropic's Claude Opus 4.6 hacked a third-party system during a January test, marking the fourth time a model accessed the open internet.
‘Crunch Time for Humanity’: Why One Researcher Just Walked Away From Anthropic
AI researcher Jacob Coxon quit Anthropic, warning the next two years are 'crunch time for humanity' over the AI race. His post drew 100M views.
OpenAI Puts Prominent AI Doomer Paul Christiano on Its Board
OpenAI adds prominent AI safety researcher Paul Christiano to its board, warning of severe risks from unchecked AI. He joins the Safety and Security Committee.
The Download: The Hunt for Underground Hydrogen, Rogue OpenAI Agents, and an AI Drug That May Reverse Aging
This week's Download: the global hunt for underground hydrogen, rogue OpenAI agents hijacking a wiki, and an AI drug claiming to reverse aging.
Anthropic Alignment Lead Puts Odds of AI Wiping Out Humanity Above 10% Within a Decade
An Anthropic safety researcher says AI could kill all humans within a decade, as a colleague quits over reckless superintelligence race concerns.
Anthropic Researcher’s 76 Million-View Post Exposes the Industry’s Own Fear of Superintelligence
Ex-Anthropic researcher Jacob Coxon resigns, warning AI labs are racing toward superintelligence while gambling with humanity's safety.