OpenAI employees warned about operational security lapses for years

OpenAI employees had warned about operational security lapses for years before the company disclosed a recording device planted in an executive conference room, according to five people familiar with the matter. Despite raising concerns internally, some staffers felt management prioritized keeping up with competitors over security.

OpenAI has begun to address these issues, appointing Thibault Sottiaux to lead an internal initiative to secure its internal systems. However, the company's monorepo, where most of its code is stored, has been criticized for being too complex and having lax security controls, potentially allowing outsiders to access sensitive information.

The Justice Department recently arrested a Saudi national for installing a hidden audio recording device in an OpenAI executive's conference room, highlighting ongoing security concerns.