The Pulse
Meta Parts Ways With Employees Hired From Virtue AI
Meta says it is letting go of employees hired from AI security startup Virtue AI, four months after the group joined in June. The company attributed the split to clashing work styles and said its AI safety focus continues.
AI.info Team ·
Meta is parting ways with employees it hired from AI security startup Virtue AI, Meta spokesperson Andy Stone told Semafor on October 2. Stone attributed the split to clashing work styles. He said Meta Superintelligence Labs remains focused on AI safety, alignment and frontier risk.
“Unfortunately, the arrangement didn’t work out as planned,” said Andy Stone, a Meta spokesperson.
Reports describe a team hire, not a company purchase
Axios reported on June 25 that Meta Superintelligence Labs was hiring Virtue AI co-founders Bo Li, Dawn Song and Sanmi Koyejo, along with other members of the startup’s team. Virtue AI’s team page also lists Carlos Guestrin as a co-founder, making clear that the three named in Axios’s report were not the company’s only founders. The terms of the arrangement were not disclosed. Axios described Meta bringing in a group of employees, not buying Virtue AI as a company.
Semafor reported that the employees were being let go four months after joining Meta in June. Its account did not name the people leaving, and said the Virtue team did not immediately respond to requests for comment.
The hires brought AI security experience
Axios said Li and Song would join Meta Superintelligence Labs, while Koyejo would report to Rob Fergus, head of Meta’s Fundamental AI Research lab. Its report described the hires as including other members of Virtue AI’s broader team as well.
Axios described Virtue AI’s work as enterprise AI security tools spanning automated red teaming, runtime guardrails and AI governance. Those areas concern testing AI systems and setting protections around their use. The reports do not say which specific projects the incoming employees worked on at Meta, or whether other Meta teams will take over any work connected to them.
Semafor reported that Virtue AI had previously done work with Anthropic, OpenAI and the Commerce Department’s National Institute of Standards and Technology. It identified the company as an AI security startup. Neither report describes a purchase or shutdown of Virtue AI.
Meta cites work-style differences
Stone’s statement gives Meta’s explanation for the separation but does not detail what the work-style differences were. He said Meta Superintelligence Labs remains focused on AI safety, alignment and frontier risk; neither report specifies how the departures will affect those efforts or the lab’s organization.
Axios reported that Meta was hiring the Virtue AI team amid wider concern about AI safety and to build up security work around AI agents. The June report said Li and Song would report to Nat Friedman, a Meta Superintelligence Lab executive, while Koyejo would report to Fergus. It did not set out an end date for their work at Meta or a later departure timetable.
The accounts establish that Meta hired Virtue AI founders and other employees in June and, four months later, said the arrangement had not worked out as planned. They do not identify everyone leaving or describe a purchase or shutdown of Virtue AI.