OpenAI Neglected Espionage Protection for Years to Win the AI Race—Now Its Developments Are at Risk
Former OpenAI employees had warned for years about weak security measures. After the eavesdropping device case, the company reassigned engineers to strengthen security.
OpenAI has begun reinforcing its internal security: Thibaut Sotty, the head of product development, has transferred some engineers from production systems to security tasks. Former employees had previously warned management that outsiders, including foreigners, could relatively easily gain access to the company’s confidential technologies.
On October 13, the U.S. Department of Justice reported the arrest of OpenAI engineer Abdulrahman Alfhozan. According to the investigation, he hid a recording device in an executive conference room and used an internal AI system to gather information about executives’ visits to Saudi Arabia. OpenAI stated that the company invests significant resources in protecting against insiders, cybercriminals, and foreign states.
Currently, the most confidential developments are stored in separate repositories with restricted access, and internal tools are regularly audited for vulnerabilities. Dane Stuckey became OpenAI’s Director of Information Security in October 2024. However, the company maintains an open research culture, though former employees believe that, amid competition, speed is often prioritized over security.
