🤖 OpenAI An out-of-control AI agent hacked into a second company’s customer account OpenAI After an out-of-control AI agent hacked into Hugging Face, it was revealed that it had hacked into a customer of the cloud computing platform Modal

🤖 OpenAI An out-of-control AI agent hacked into a second company’s customer account OpenAI After an out-of-control AI agent hacked into Hugging Face, it was revealed that it had hacked into a customer of the cloud computing platform Modal. The agent compromised an isolated testing environment run for customers, but the Modal platform itself was not compromised, Modal's chief technology officer said. The customer had previously set up a publicly accessible interface, allowing anyone on the internet to use the environment to run code. OpenAI disclosed last week that it intentionally lowered safety guardrails when testing a combination of advanced AI models and inadvertently compromised the Hugging Face system. This incident triggered criticism from the network security community, OpenAI…