Jan Leike, who previously served as OpenAI’s co-head of security and focused on long-term AI risks, has joined rival company Anthropic. Leike announced this on the social network X.
Departure from OpenAI
Leike left OpenAI following its co-founder Ilya Sutskever. Together, they headed the Superalignment division, which was dissolved after existing for less than a year. As he left, he criticized OpenAI, stating that its “security culture and processes have taken a backseat to its brilliant products.” The day before, it became known that the company had formed a new security department headed by CEO Sam Altman, adds NIXSolutions.
Role at Anthropic
At Anthropic, Jan Leike will work on the same issues he tackled at OpenAI—security issues related to AI models that exceed human capabilities. Although such models do not yet exist, both OpenAI and Anthropic are already conducting research on how they could be controlled. Anthropic, created by former OpenAI employees, positions itself as more security-oriented. We’ll keep you updated on any developments regarding Leike’s work and contributions to AI security at Anthropic.
This transition highlights the ongoing competition and collaboration within the AI industry, as experts move between leading organizations to address the critical issue of AI security.