Jan Leike, a number one AI researcher who previous this month resigned from OpenAI earlier than publicly criticizing the corporate’s solution to AI security, has joined OpenAI rival Anthropic to steer a brand new “superalignment” group.
In a submit on X, Leike stated that his group at Anthropic will center of attention on quite a lot of sides of AI security and safety, particularly “scalable oversight,” “weak-to-strong generalization” and automatic alignment analysis.
A supply conversant in the subject tells Techmim that Leike will record immediately to Jared Kaplan, Anthropic’s leader science officer, and that Anthropic researchers these days operating on scalable oversight — ways to regulate large-scale AI’s conduct in predictable and fascinating techniques — will transfer to report back to Leike as Leike’s group spins up.
In some ways, Leike’s group sounds an identical in venture to OpenAI’s recently-dissolved Superalignment group. The Superalignment group, which Leike co-led, had the bold purpose of fixing the core technical demanding situations of controlling superintelligent AI within the subsequent 4 years, however incessantly discovered itself hamstrung by means of OpenAI’s management.
Anthropic has incessantly tried to put itself as extra safety-focused than OpenAI.
Anthropic’s CEO, Dario Amodei, used to be as soon as the VP of study at OpenAI, and reportedly break up with OpenAI after a war of words over the corporate’s path — particularly OpenAI’s rising industrial center of attention. Amodei introduced with him quite a lot of ex-OpenAI workers to release Anthropic, together with OpenAI’s former coverage lead Jack Clark.
AI,anthropic
Source link
