In the spring of 2023, a hacker broke into an internal forum at Open AI where employees discuss technical details about the company’s technologies.
Open AI was hacked over a year ago but chose to keep quiet about the breach, the New York Times reports. The breach took place in an internal forum where employees discuss the development and technical details of the company’s various AI models.
The system is independent from other internal systems and the hacker did not come across any login information to other parts of the company. Since the forum does not contain any customer data, the company chose not to disclose the breach to the public, but only to employees and the board. The company also did not contact the FBI or any other authority.
Not everyone agreed with that rationalization. One executive, Leopold Aschenbrenner, criticized the company’s security measures, which he argued left the company open to major intrusions by state actors. He was later fired for leaking internal secrets, which he claims was politically motivated. Open AI says his firing is not related to the security criticism, and that management simply came to a different conclusion.
The New York Times writes that Open AI does not believe its AI algorithms pose any risks to national security, and is supported by others in the industry. Daniela Amodei, CEO of Anthropic, says, for example, that today’s AI systems probably cannot be used to cause any major damage. Meta has made its models, led by Llama, open source and is not alone in that.
2 Comments
Recommended Comments
Join the conversation
You are posting as a guest. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.