Meta Superintelligence Labs is hiring three co-founders and other team members from AI security startup Virtue AI to strengthen its safety work around AI agents, according to Axios. Co-founders Bo Li, Dawn Song, and Sanmi Koyejo, all leaders in AI security, are joining alongside members of Virtue's broader team, which builds enterprise tools spanning automated red teaming, runtime guardrails, and AI governance. Li and Song will report to Meta Superintelligence executive Nat Friedman, while Koyejo reports to Rob Fergus, head of Meta's FAIR research lab, with the terms undisclosed. The move follows a swell of AI-safety scrutiny after the release and pullback of Anthropic's latest models, and continues Meta's pattern of poaching discrete teams rather than acquiring whole companies, as it did with the founders behind agent startup Dreamer in March.
