So in case you haven't heard about it, Microsot designed an AI that learns and evolves from talking with people in twitter.
It soon turned into a racist genocidical machine by troll comments and was pulled out.
But what I find really funny about this is that it means the Megaman X series turned out kinda prophetic. X only turned out a good robot because he spent decades in a bunker "learning" how to be good, while basically every other AI in the series designed to interact with humans eventually degenerates into "Kill the filthy meatbags!"
That's also why "true AIs" are really overrated. If they can learn anything, there's a lot of wrong stuff out there to learn.
Actually, if I remember correctly, Dr. Wright basically pre-programmed X before sealing him away in a time capsule. The reason X is immune to the Maverick virus is simply because he's an older architecture that can't interact with it.
That's incorrect by at least three reasons:
-Most reploids were built by reverse-engineering X.
-Zero was built around the same time as X, and Zero's the origin of the Maverick virus.
-In the Megaman Zero time period, the first game's final boss is a carbon copy of X that was built after the original X went missing, except they skipped the whole "spend decades learning the difference between good and evil" bit, and thus the copy X used its superior power to basically become a psycopath robot dictator (leading to the period's premise of scavenging Zero because he was pretty much the only bot who ever could go toe to toe against X). It's one of my favorite boss fights in Megaman games, since copy X has a dynamic AI that can use any of the original X weapons from the previous games, so it can be pretty different duel each time you fight it.