Re: This - read up on Elon Musk's comments and the consortium he's leading to try and ensure that this doesn't happen. It's a very real problem - not just some alarmist nonsense drummed up to scare people. I've spoke with a PhD who majored in AI and cognitive recognition, and he was as concerned about it as Musk is. "It's going to happen, we just don't know when" and "It's fine for us to try and control it in civilised societies, but think about the underground/3rd world experimenters who may unleash something that we can't control". He likened it to the unregulated environment that exists around cloning humans.
And here's the simple reason: When AI 'learns' it has the ability to iterate thousands, if not millions of times in a short period of time, so that it can get something 'right'. AI can become proficient in a fraction of the time and will master something in a moments, that would normally take us many years. Musk, and others have stated that because of this fact alone we must tread very carefully when we create, and then deploy AI in order to do certain things. The argument is that as soon as AI starts going rogue, it goes very rogue, very quickly - from 'oh wow, my AI-powered robot burnt my toast' to 'oh cr@p it just tried to kill me in my sleep!' in hours, or possibly even minutes.
Yes, yes... There are movies about this, and "....the threat isn't real" and "But we're in control, as we program the machine". Well. Maybe not so much... This is the first time in history we have attempted to create something that doesn't just extend our cognitive abilities... But, we are willing engaging with tech that can supersede our cognitive ability - And that's dangerous. Because as it soon as it realises it can communicate in a way that we can't understand - think the current problems we have with quantum computing - than it will quickly be able to invent other ways to mentally, emotionally, physically, financially, and socially dominate us completely.
It's worth reading up on. Absolutely fascinating. I'm not a paranoid alarmist. I want to be informed, so that I can try and reason with our new AI Overlords... ;p
