Meta Superintelligence Labs has hired the founders and several team members from AI security startup Virtue AI, according to a report based on an internal memo viewed by Axios.

The move adds more security expertise to Meta as the company works to strengthen protections around its AI systems and agents. The hires come at a time when AI safety has become a bigger issue across the industry, especially after Anthropic recently released and then pulled back one of its latest models amid renewed scrutiny.

Virtue AI co-founders Bo Li, Dawn Song and Sanmi Koyejo are among those joining Meta, Axios reported. The broader Virtue AI team is also moving over, including employees who work on enterprise security products such as automated red teaming, runtime guardrails and AI governance. The financial terms of the arrangement were not disclosed.

In the memo, Meta said that safety and reliability are central as it rolls out AI products to a large user base and develops more capable agents. The company has been emphasizing security as its models and products become more powerful and more widely deployed.

The new hires will be split across Meta’s AI organizations. Li and Song will report to Nat Friedman, an executive at Meta Superintelligence Labs. Koyejo will report to Rob Fergus, who leads FAIR, Meta’s fundamental AI research group.

The addition of Virtue AI talent reflects a broader push by major AI companies to focus on agentic security. As AI systems take on more autonomous tasks, labs are paying closer attention to how those systems can be tested, constrained and monitored. That concern has intensified as regulators and policymakers examine which models may face added scrutiny.

The reporting also notes that the White House may consider pulling certain AI models, underscoring the rising pressure on developers to prove their systems are safe. At the same time, OpenAI expanded access to its cybersecurity tools this week without drawing the same level of attention, according to Axios.

Meta has been using targeted recruiting to bolster its AI efforts. In March, the company also hired the founders and team behind Dreamer, another startup focused on helping users build agents. That pattern suggests Meta is not only chasing top technical talent, but also building out specialized groups that can support its broader AI strategy.

For Meta, the latest hires appear aimed at reinforcing trust and security around future products. As competition in AI intensifies, the company is making clear that safety work is becoming a bigger part of its superintelligence push.