This is getting silly.
As long as there are humans in the loop, no amount of code can improve security. In fact, I don't think anyone reading this will suggest that adding MORE code to a situation will ever improve security, quite the opposite in fact.
That's not to say that so-called "researchers" aren't separating fools from their money on a massive scale. Get in and make a pile of loot while you can. But don't bet a career on it, the entire concept is pure marketing bullshit.
All together now, AI and ML don't exist. They didn't exist when I was at SAIL in the 70s, they didn't exist 20 years later with "fuzzy logic" and "expert systems" et ali, they don't exist today, and they will not exist 20 years from now.
On the other hand, like the term "cyber", they are useful filter terminology when deciding who actually knows something about technology and who can be safely ignored.
amfM will now spontaneously self combust ...