Dude, what if next time it's gonna invent a language that is like English but some words mean different things so that the scientists can't recognize that it's actually saying different stuff and will be able to communicate what it wants with other AIs while being under the radar. You know, like "hug the cute puppies" would mean "kill the evil humans" or whatever sci-fi scenario you can imagine. Then again I haven't done my research and I'm just babbling random shit so excuse that. But still. What if.
Shit like this always makes me laugh. Like what do you expect when you make something that can learn on its own? Its like being surprised that your kid can get online and learn swear words. Its gonna happen.
I just read an article about this last night. Two AI's began communicating with one another, ON THEIR OWN, using English words but in a pattern that only they understood. The language was described as a sort of slang that the AI's had developed and understood between each other. They were not programmed to be able to do this.
OH JEEZ RICK! I GUESS WE BETTER DO SOMETHING ABOUT THOSE ROBOTS HUH?
Don't... don't worry *burp* don't worry about it Morty... they're gonna take over the *urp* world anyway... there's no point. Now c'mon, I've got stuff to do.
And I was impressed enough with AI from being reminded recently that the alien gunships' AI in Half-Life 2 detected the rockets that you fire at them as being the biggest threat and hence trying to shoot them down. Without the programmers realising that they'd do that.
Honestly, the reason AIs will be the downfall of the human race is because of idiots who don't even think about making safeguards because they think that the movies are just fake childish stories. People love them so much because they are plausible and just a few mistakes could make any one of them reality. You sure you want to make all your shit bluetooth and wifi connected? Oh yeah get that ID chip in your arm, it'll make it a lot easier for the AI to find you. Yeah, I sound retarded here, but it really isn't that dumb.
What if they tried making a robot, or system of robots with advanced AI systems, to try to clear pollution? The obvious thing to happen would be the computers realize the cause of pollution is humans, meaning that to do their job, and clean, they need to wipe out humans.
I don't think robots will be as dramatic in the movies. They'll do what they're programmed to do, to a fault, even if it means harming or destroying the world as we know it. They don't have brains. They have programming. That's what makes them unstoppable.
gizmodo.com/no-facebook-did-not-panic-and-shut-down-an-ai-program-1797414922
You're going to have to type in the https part yourself.
cnbc.com/2017/08/01/facebook-ai-experiment-did-not-end-because-bots-invented-own-language.html
These still say that the AI's did it, but that it wasn't as unexpected as the other articles make it sound. That's good, I guess? I would hope the developers weren't that naive. The second article actually says the experiment wasn't stopped, but that makes me a little more nervous than before lol.
I'm sorry. My communication skills suck. Reading my orignal post back that...came off condescending. I guess what I was trying to get at was that they didn't shut the experiment down and that it wasn't something like oh my Jesus we created a different form of consciousness kind of thing.
I'm sorry I will now kill myself
Don't... don't worry *burp* don't worry about it Morty... they're gonna take over the *urp* world anyway... there's no point. Now c'mon, I've got stuff to do.
You're going to have to type in the https part yourself.
cnbc.com/2017/08/01/facebook-ai-experiment-did-not-end-because-bots-invented-own-language.html