Meta parts ways with the Virtue AI safety team it brought in four months ago, Semafor reports

Semafor reported on 2 October that Meta is parting ways with staff it brought in from Virtue AI, an AI security startup, in June. They include Bo Li, Dawn Song and Sanmi Koyejo, all founders of Virtue AI, and others; Semafor gives no headcount and no deal terms.

Meta spokesperson Andy Stone told Semafor: “Unfortunately, the arrangement didn’t work out as planned”. Semafor reports that the reason was clashing work styles, and that Stone said Meta Superintelligence Labs remains focused on AI safety, alignment and frontier risk.

Virtue AI sells an enterprise AI safety platform; its own team page lists Li as founder and chief executive, Song as a founder and board director, and Koyejo as a founder and chief AI officer. Semafor's report is about the people who joined Meta, and how their move relates to the startup itself is not visible from here.

This is one outlet's report, with Meta's comment, and no statement from the researchers was found. It lands on the same day OpenAI confirmed it had fired three researchers and that a safety and transparency staffer had left. The movement of safety people out of the big labs is now a weekly story, and it matters because these are the people who design and publish the tests the public relies on.

• Reported Meta is parting ways with the Virtue AI staff it brought in in June, including Bo Li, Dawn Song and Sanmi Koyejo. Semafor
• Claimed Meta says the arrangement did not work out as planned and that Meta Superintelligence Labs remains focused on safety, alignment and frontier risk. Meta (Andy Stone), to Semafor