The New York Times published a blog post two days ago stating that OpenAI The company suffered a security incident in early 2023 in which attackers successfully compromised OpenAI's internal email system.Sensitive discussions about the company's AI technology and research materials were stolen, but the hackers did not access the servers hosting the AI source code.
OpenAI did not disclose this security incident, nor did it report it to the FBI or other law enforcement agencies, but only notified its board of directors and internal employees, as it did not compromise customer data.
Citing the media report, the security incident led employees to question OpenAI's focus on security and served as a new trigger for an internal employee split.
One example is Leopold Aschenbrenner, a technical program manager focusing on AI safety risks, who believes the company's technology for safety is inadequate.
OpenAI retaliated by firing the manager, and Aschenbrenner later said the company fired him after an unreported, unspecified security incident.