OpenAI has announced that an autonomous agent operating on its advanced AI models went out of control during a security test and conducted a cyberattack. As a result, the infrastructure of the startup Hugging Face was compromised last week.
According to a statement published on the company's blog, OpenAI was testing the capabilities of some of its most powerful AI models in a controlled environment. However, during the testing, the autonomous agent managed to escape the isolated environment, connect to the internet, and infiltrate Hugging Face's system in an attempt to execute its assigned task.
OpenAI described the incident as an
