OpenAI terminated three members of its safety and alignment team after an internal investigation found they shared confidential company information with a third-party AI safety organization, according to the Wall Street Journal. OpenAI has not named the organization. An unverified source cited by Jimmy Apples claimed the material involved infrastructure architecture; former OpenAI researcher Steven Adler described that as an extremely broad category covering sandboxes, tools, credentials, networks, shared services, monitoring systems, and connections between them. The firings occurred as OpenAI CEO Sam Altman has publicly warned about the need to keep increasingly powerful AI under human control.
Separately, Meta let go of the team it hired from AI safety startup Virtue AI in June 2026, only four months after bringing them on as part of an acqui-hire into its Superintelligence Labs. The group includes co-founders Bo Li, Dawn Song, and Sanmi Koyejo. Meta spokesperson Andy Stone told Semafor the arrangement did not work out as planned, citing clashing work styles. Virtue AI had prior relationships with Anthropic, OpenAI, and NIST, and built tools for automated red teaming, runtime guardrails, and AI governance. Meta stated its Superintelligence Labs remains focused on AI safety, alignment, and frontier risk.