The more advanced AI becomes, the more attacks on legacy systems we will see happen. And those people who are not using AI to update and upgrade their security are the ones at fault here. How do they expect to go toe to toe with current AI technology? Do they really believe that their human written code has no exploitable parts? Or do they just arrogantly assume it is safe?
Way too many service providers are too comfortable with being too greedy to AI-proof their systems. We need some legislation or some kind of mandatory new special AI insurance for these companies. They do not really seem to care about their customers otherwise.
Humans haven't written a line of code for years now, they are just in there to double check it. Even best penetration stress test for it is done with AI.
And no one is assuming anything has ever been "safe". Just windows alone comes with new security patches all the time, and we can only hope someone wasn't fast enough to exploit them before they got patched. Exploits exist and everyone who knows tech knows it, most of them haven't just been found yet.
But by all means, offer an alternative to this. I believe that some sort of AI insurances already exist, but we can't just put words together and pretend it's a thing we can demand. So much has changed in short time, so who is supposed give these new AI insurances and with what terms? Those terms and threats can change in a day.
