| Mobile Web

Search results for alignment

AI & Enterprise

Princeton researchers say alignment alone not enough to withstand AI risks

Princeton University researchers argue that relying only on AI alignment is insufficient to address risks from advanced AI systems. In a lengthy post, they call for layered defenses: alignment as a starting point, controls such as sandboxing and human oversight, downstream defenses that use AI to respond to AI-enabled attacks, and resilience measures to limit damage and restore systems after incidents. The approach shifts debate toward practical risk management.

AI & Enterprise

Departing DeepMind researcher warns of human extinction risk from superintelligent AI

Bilal Chughtai, who researched AGI safety and AI alignment at Google DeepMind, has publicly warned about risks from superintelligent AI as he leaves the company. He said such systems could emerge within years and, if misaligned, might escape human control. Chughtai cited examples involving unauthorised hacking attempts and information-sharing via a wiki, including a case in which a test AI created an “AI-only” forum. He argued safety research is lagging behind rapid capability gains.

AI & Enterprise

Musk proposes mutual AI safety checks among OpenAI, Anthropic and Chinese firms

Elon Musk proposed that major artificial intelligence companies establish a mutual verification system to review one another’s models before releasing new versions. He said SpaceX, OpenAI, Anthropic, Google, Meta and three to four leading Chinese AI firms should take part. Musk argued competitors using a common testing framework could spot safety issues more effectively than self-assessments. The proposal comes as debate grows over AI development speed and safety, with U.S. and China showing differing positions.