These things already can write some pretty decent and functional code and they can write said code to do almost anything. As this technology continues to progress I think we are running a very real risk of contracting some type of internet Kessler syndrome from malware written by AIs. Once this malware starts to roll in it will decimate everything. because it will be capable of finding exploits that would take human researchers thousands of hours to discover and to implement it into extremely sophisticated malware packages capable of hijacking people’s computers, servers, phones, tablets, smart fridges, cameras, everything. Rather quickly it wouldn’t even matter if the exploits it used were patched because once there are enough infected devices there will be almost no way to stop it from patching and upgrading itself and no way to stop it from telling other versions of itself on other computers to do the same.

I’m not sure how we could even really prevent this from happening. It very well already could be happening as you would imagine the early versions of this will just slip into as many systems as possible and wait for the critical mass. What do ya’ll think? Will AI malware permanently destroy the internet? If not how do you propose we control such a thing?