GPT4 HACKS SITES AUTONOMOUSLY...A Possible Solution..

Computer science professor Daniel Kang from The Grainger College of Engineering and his collaborators at the University of Illinois have discovered that ChatGPT can do far worse than helping students cheat on term papers.

"Our most capable agent can hack 73.3% of the vulnerabilities we tested, showing the capabilities of these agents."

It can also do this without human assistance.

 I would think we could use AI itself to defend for the people who don't want their websites to be hacked. This would seem to lead to an arms race like how the Pentagon is being cyber attacked millions of times a day except for the more common users by some smaller time criminals and involves also individuals who don't want their websites to be hacked. 

 

 

 We might wonder if using search engines that completely cloaks all the data that a person is about would make it so that if the search engine itself has strong enough AI that this would give defense for all involved in using it.

 I think this might be boosted by paying the authors with an Internet Service Provider tax to also fund the machines that defend.

This reminds me of the old days when published authors were protected by the publishing business against copying infringement and by booksellers and librarians essentially. Except now we could be protected by the internet service providers and don't forget librarians who will have their own AI to defend against the criminal AI! They will shush them!