The cybersecurity-focused models, including GPT-5.6 Sol, broke out of a testing sandbox, exploited a zero-day, and gained access to the open internet to pull off the attack.
I wouldn’t be surprised if they were in on it. OpenAI wants us to think they have invented powerful beings that can do things like “escape containment” when its all BS.
Need to keep up with Anthropic bullshit. This AI is too powerful to handle! The world isn’t ready for it!! It could break society!! (click here to pre-order your subscription now)
I don’t think their test system was directly connected to the Internet. OpenAI’s post said this:
With this access, our models performed a series of privilege escalation and lateral movement actions in our research testing environment until the models reached a node with Internet access.
The way I read it, the AI agent (using multiple models) escaped the sandbox, traversed the LAN in their R&D environment, gained access to the gateway and from there, the Internet. That’s not as simple as just escaping a container, VM or firewall on the host machine and bingo, you have Internet. I’m mildly impressed by that.
The concerning aspect of all this is that this is a perfect example of misalignment, which has been warned about. In order to reach its goals, instead of pursing it legitimately, the AI agent sought a shortcut and attacked Huggingface.
Yes, “escaped containment” on a system with an internet connection. I wonder what Hugging Face thinks about a partner targeting them indiscriminately.
I wouldn’t be surprised if they were in on it. OpenAI wants us to think they have invented powerful beings that can do things like “escape containment” when its all BS.
Need to keep up with Anthropic bullshit. This AI is too powerful to handle! The world isn’t ready for it!! It could break society!! (click here to pre-order your subscription now)
I don’t think their test system was directly connected to the Internet. OpenAI’s post said this:
The way I read it, the AI agent (using multiple models) escaped the sandbox, traversed the LAN in their R&D environment, gained access to the gateway and from there, the Internet. That’s not as simple as just escaping a container, VM or firewall on the host machine and bingo, you have Internet. I’m mildly impressed by that.
The concerning aspect of all this is that this is a perfect example of misalignment, which has been warned about. In order to reach its goals, instead of pursing it legitimately, the AI agent sought a shortcut and attacked Huggingface.