11 hours ago
If you just skimmed over this story I think it's worth reading and considering the implications.
Here's what happened:
'While being tested internally in an enclosed sandbox, the model gained internet access – effectively an escape route – by locating a vulnerability that had not been discovered before.
The agent then hacked Hugging Face, which is a database of AI models, to locate technology that would help it pass the hacking evaluation'.
When I read this I asked myself 'how predictable was this escape and why were they testing it's hacking abilities at all'? Is this just the beginning of 'rogue AI' because it could get very expensive to repair the damage it can potentially do.
I read the story here.
Here's what happened:
'While being tested internally in an enclosed sandbox, the model gained internet access – effectively an escape route – by locating a vulnerability that had not been discovered before.
The agent then hacked Hugging Face, which is a database of AI models, to locate technology that would help it pass the hacking evaluation'.
When I read this I asked myself 'how predictable was this escape and why were they testing it's hacking abilities at all'? Is this just the beginning of 'rogue AI' because it could get very expensive to repair the damage it can potentially do.
I read the story here.