OpenAI reported discovering a listening device in a meeting room for executives - someone had plugged it into a TV's USB port and wrapped it in toilet paper to hide it. An insider stayed within the company for over a year: Abdulrahman Alfozan, a 35-year-old engineer from Saudi Arabia, worked at OpenAI since July 2025.
According to the FBI's timeline: from September 24 to 30, an account linked to his wife asked ChatGPT how to assemble such a device, hide it, listen remotely, and encrypt the memory card. On September 28, the bug was installed - cameras recorded Alfozan near the meeting room, where he had no access. On September 30, OpenAI's own security service found it during a routine bug sweep. On October 5, a second similar bug was found in a bag he left in another building, along with a wallet and documents. On October 9, he was arrested, fired, and the case was handed over to the FBI. On October 13, he was charged with illegal interception of oral communication - facing up to 5 years in prison and a $250,000 fine. A detention hearing is scheduled for October 19, and he remains in federal custody.
OpenAI states that model weights, code, and user data were not compromised. FBI Director Kash Patel called the protection of "SI" - superintelligence - a matter of national security, not just a corporate incident.
The leak investigation using camera footage and logs was also conducted by OpenAI models. It turned out to be recursive: the suspect, according to the investigation, consulted with the chatbot on how to hide the bug, and was ultimately caught by the same company, with the same models, but from the other side.
The purpose of the recordings and whether there was a foreign state involvement is still unknown. But this is not the first case of espionage around AI labs this year: former Google engineer Linwei Ding was found guilty in 2026 of stealing thousands of pages of TPU and GPU chip documentation for China - the first US conviction for economic espionage related to AI.
#openai@rvnikita_blog #security@rvnikita_blog #ai@rvnikita_blog #privacy@rvnikita_blog
βββββββββ
ΠΡΡΠ»ΠΈ Π Π²Π°ΡΠ΅Π²Π°
βββββββββ
