Former OpenAI employees raised concerns about the company’s security practices even before ChatGPT launched in late 2022, according to five people familiar with the matter, taking their worries to senior leaders and the board, with some leaving those conversations believing management cared more about keeping pace with rivals such as Google and Anthropic than about security. The concerns followed OpenAI’s October disclosure that a recording device had been found planted in an executive conference room; on October 13, the US Justice Department said federal agents had arrested a Saudi national and charged him with installing a hidden audio recording device, concealed in toilet paper, in an executive conference room, and Abdulrahman Alfozan, 35, was an OpenAI engineer at the time. A court filing says he also used an AI tool built for OpenAI staff to find internal documents about executives’ visits to Saudi Arabia and the officials most likely to attend meetings with Saudi ministers, and that he was in regular contact with a Saudi royal who tried to recruit him for a new AI venture in the kingdom. Much of OpenAI’s code sat in a single repository, known as a “monorepo,” which let anyone with access view details of projects unrelated to their work, the five people said, and three people said OpenAI had not run regular security reviews of internal tools such as the sandboxes where researchers stress-test new models. OpenAI disputes parts of this picture, saying its most sensitive work is held in separate repositories with strict access controls, that the monorepo does