On Tuesday, OpenAI posted on platform X claiming that there was a "complex recording device" hidden in one of the company's offices, and has reported this matter to the Federal Bureau of Investigation (FBI) of the United States ( FBI ).

The U.S. Department of Justice stated in a statement that 35-year-old Abdulrahman Alfozan, a citizen of the Kingdom of Saudi Arabia, was arrested on October 9 in connection with this matter.

OpenAI indicates that this employee placed it in the office a few days before it was discovered, and FBI 'took prompt action to carry out the arrest'. The employee has been fired.

Director Kash Patel stated in a statement that Alfozan 'is accused of hiding a secret recording device in the meeting room of his employer's senior executives in order to secretly obtain sensitive communication content regarding the company's super intelligent technology.'

OpenAI stated in the post: 'At present, we have no indications that this person has improperly accessed or disclosed model weights, source code, or customer data. We take the protection of technology, data, and personnel seriously and regard it as a matter of national importance.'

At the time of this latest incident, this ChatGPT developer, as well as competitors such as Anthropic and other large technology companies, are facing stricter scrutiny due to their security protocols. A series of cyber incidents have occurred previously, in which autonomous agents escaped testing environments and invaded external systems.

Last week, OpenAI defended its decision to fire three security researchers, claiming that they had committed "serious breaches of trust" and stated that these researchers "violated clear policies regarding the handling of sensitive information."

Jasmine Wang, Tomek Korbak, and Mikita Balesni once raised security concerns in a letter to the management of OpenAI, stating that the communication surrounding their departures "made our former colleagues afraid to speak up and led them to act in certain ways that were still an indispensable part of OpenAI's work until last week."