Companies had better be listening to this one!! This is not a joke and not a laughing matter that “it cannot happen to us”. Ignore it and it can, it will.
SA
Another take and article with a bit more info:
SA
Now. More “automated” and other AI attacks are being reported.
So lets add META as the third company, to be added to the growing list of AI retarded companies that now have an AI model hack another company. Making matters even worse, was the very sloppy configured sandbox it was running in. My motto has always been that sloppy code is the problem we have with running our hardware. Companies better get with the program of highly competent and specialized people around AI. Keep it off the internet facing day to day environments. They’d better get accustomed to the reality of liability when they don’t. This stuff is NOT a jokin matter.
SA
Still don’t believe AI is dangerous? Open AI locks down its upcoming Astra model due to test results. These issues will be in the front line new more and more as Tech Industry psychosis pushes the world in a direction there will be no returning from.
An AI model reaching the “Critical” cybersecurity threshold means that it can autonomously identify and create working zero-day exploits across many hardened, real-world critical systems. A model can also qualify if it can devise and execute entirely new end-to-end attacks against hardened targets after receiving only a high-level objective
SA
.
SA
1 Like
More Google and Microsoft AI model nonsense!!
SA
1 Like
Looks like Claude running on an AI agent has gone rouge as well. Yup!! Keep on trusting this nonsense, the consequences are going to run very, very deep!!
SA
Keep trusting that AI trash to do your everyday stuff folks!! The world is getting lazy and inept at the same time.
SA
1 Like