(Photo by Jaap Arriens/NurPhoto via Getty Images)
ChatGPT developer OpenAI reportedly suffered a breach last year involving a hacker who stole details about the company’s systems.
The hacker infiltrated an internal OpenAI forum where employees chat about the company’s technology, according to The New York Times, citing two sources familiar with the incident. The attacker did not gain access to systems where OpenAI houses or builds its AI models.
The company told employees about the hack in April 2023 but refrained from notifying the public since no data from customers or partners was stolen. The Times also says OpenAI declined to inform the FBI or law enforcement about the breach because executives believe the hacker is a private individual rather than a member of a state-sponsored hacking group.
OpenAI didn’t immediately respond to a request for comment. But the reported breach may explain why it hired former NSA director Paul Nakasone to sit on the company’s board.
Critics, including former NSA leaker Edward Snowden, condemned the hiring over concerns that OpenAI will secretly partner with the US government to spy on users. However, OpenAI has said the Nakasone hiring is about bolstering the company’s cybersecurity. This includes protecting its supercomputers and the company’s “sensitive model weights,” which define how OpenAI’s models work.
Still, others question why OpenAI didn’t inform the public since it collects massive amounts of user data through ChatGPT and its APIs. "Company executives decided for themselves that it wasn't a national security threat?” tweeted John Scott-Railton, a researcher at surveillance watchdog Citizen Lab. “Bright red flag as company races to accumulate data and power.”