Did a Hacker Breach OpenAI’s Internal AI Discussions? Implications for Security

OpenAI Internal Messaging System Hacked

According to a New York Times report, a hacker infiltrated OpenAI’s internal messaging system early last year, obtaining employee discussions about the company’s latest artificial‑intelligence developments. The breach did not reach core code, but it sparked concerns about foreign adversaries stealing sensitive AI secrets.

The company disclosed the incident to staff and the board in April 2023 but kept it private, stating that no customer or partner data was compromised and that the attacker acted independently, not as a state‑backed actor. OpenAI did not notify law‑enforcement agencies such as the FBI.

Employees worried that nations like China or Russia might exploit the stolen information, raising national‑security alarms. Former technical manager Leopold Aschenbrenner criticized OpenAI’s security measures in a memo and was subsequently dismissed, a move the firm says was unrelated to his criticism.

OpenAI spokesperson Liz Bourgeois acknowledged the concerns but denied that they led to Aschenbrenner’s departure, emphasizing the company’s ongoing work on safe AGI. The episode adds to broader industry anxiety about the potential weaponisation of powerful AI and the rapid AI talent growth in China.

Recent resignations of senior staff, including chief scientist Ilya Sutskever and researcher Jan Leike, over safety issues underscore the pressure on OpenAI to strengthen its security posture. In response, the company created a safety committee that includes former U.S. national‑security officials, while calls for governmental regulation of AI intensify.