I think the cat is out of the bag now.

Regulations of AI agency and commercial services will not prevent rogue agents still causing malicious activity.

Sure, it will reduce it, but it's not a guarantee that if the big AI companies start providing only regulated, "safe" agents, noone else will use the already existing models just to keep doing that from the comfort of their own rack.

The more I think of it (although I am just a mere person with no ordinary archievements), the only viable solution is to reinvent the internet to restrict anonymous, consequence-free access, sort of "taking ownership" of your digital footprint and then whoever allows an agent to do bad stuff needs to answer for it.

It would be an incredibly difficult and international effort to pull something like that off, but without that I feel we're slowly diving into sort of a Cyberpunk 2077 scenario, where outside of local intranets noone connects to "The" Internet as rogue AI agents can kill you there.