OpenAI kept a hacker attack on its internal communications platform secret from the outside world. The incident caused quite a bit of unrest among our own employees.

The New York Times reports on a security incident at OpenAI. A hacker is said to have broken into the AI ​​company’s internal communications platform in order to read employee conversations. The intruder was unable to access the underlying systems of OpenAI’s AI models.

The incident in question occurred in early 2023 but has remained under the radar until now. OpenAI chose not to report the incident to law enforcement and nothing was shared with the outside world.

Company management reportedly did not consider this necessary, as the intruder was unable to extract any damaging secrets and was believed to have acted individually and without ties to any nation-state. Perhaps OpenAI’s management did not want to jeopardize its rapid rise in the technology industry.

Dirty laundry

The company’s own employees were informed and the incident is said to have caused considerable unrest. Employees questioned OpenAI’s security policies. If the intruder had been working on behalf of a hostile actor, sensitive information could have been leaked.

The hack was partially uncovered by a former employee named Leopold Aschenbrenner. OpenAI fired him in the spring, and since then Aschenbrenner has been airing dirty laundry about the company without batting an eyelid. In a podcast, he openly said that OpenAI’s security wasn’t strong enough to keep out intruders with bad intentions. He also claims that his firing was “politically motivated,” which doesn’t make him the first former employee to follow suit.

OpenAI denies this in a response to the New York Times and emphasizes that it is fully committed to developing secure AI technology.