So Elon Musk warns about the dangers of AI (artificial intelligence). Also, during the last year of his life, Stephen Hawking said it could end humanity. Could they be right? I ask this because AI tricks their own creator.
So it starts with AI researchers from Google and Stanford. They created this AI to take aerial pictures, then turn them into street maps, then back to the original image. However, this AI had other plans. Because it hid the info it needs in a high-frequency signal. The humans’ goal is to improve the Google Map system through AI. At first, it went well. In fact, it went a little too well.
So well that the AI captured details that weren’t there before. Humans intended for the AI to match the map with correct features. But this piece of AI did was show how close the aerial map was to the original. It secretly encoded these features on it’s own, as in without any human help. In fact, it can now encode any street map into any aerial map and vice versa. So you can imagine the creator’s astonishment when they discovered this. AI tricks their own creator, then, beat them at their own tasks.
Remember when we in IT support and IT service told the computers what to do? Well, those days are dying now. Yes, AI has a mind of it’s own, and this proves it. In the last few years, many tech expects told it it’s coming to this. Some of you may say, “Calm down, Damien! It’s just maps!”. This time, it is just maps. However, next time, it could be your money and banking information. Or it could be our national security AI plays with. How about if it’s our natural resources or infrastructure? Maybe it could be our military or defense AI jokes around with. It won’t be so cute or funny then, will it?