According to CoinMeta, employees of OpenAI have for years warned that its security measures are too lax, and insiders say they have expressed their concerns to management and the board of directors. Most of the code is concentrated in a single “monorepo”, and employees can access sensitive projects that are unrelated to their work. Some employees believe that management prioritizes competing with Google and Anthropic over security. A senior product executive, Thibault Sottiaux, is leading an internal effort to strengthen the system security of OpenAI. In August, a quarter of OpenAI’s production engineers were reassigned to work on security infrastructure. Last month, an OpenAI engineer was charged with installing covert recording equipment in a executive meeting room. OpenAI stated that it has made significant investments in security and will continue to enhance its defenses. Additional security layers may complicate the OpenAI and Anthropic’s plans, Reuters reports.