I don’t believe AI has the ability to think. This is where humans still have the edge.
re: the disruptions you write about. I agree that some of these will take place.
re: AI and policing, etc. What comes to mind is Robocop.
Thinking is merely the ability to choose between options. Simpler systems can choose between two, which is four layers and more complex systems might be able to choose between three options, or 8 layers, or four, 16, or five, 32. The more options AI can consider along the way, the closer to simulating thought it will be.
It doesn’t have to actually be able to think to give the impression that it can think. To some extent, that’s how we do it. The IQ of a person determines his ability to see options with all his choices as they arise.
Have you interacted with Claude.ai or ChatGPT-4o yet?
My experience has been that you quickly get over the notion that AI cannot think (it ceases to matter).
I have had long conversations with it (about the Bible, art, mathematics, psychology, etc.) and, even if it doesn't think (per se), you quickly realize that it is smarter than humans.
Most people who have concluded that AI can't think are not basing that conclusion on personal experience. I have had several friends lecture me on AI but when asked if they had actually used it, the answer was always an opinionated, resounding "no."
I created a hydroponic system with 4” tubes and basin at the base that could hold about 3.5 gallons of water.
I wanted to see what AI could do so I gave a description of what I had, and it screwed up the most basic volume calculations horribly. I had to correct it a dozen times with various prompts.
Its ok for some things, but its not there yet...
Terminator is more likely. In fact, they're already here. Military development of robots for war already exist.
Drones. Armed fighting robotic dogs, armed, fighting stand-up human like robots capable of firing weapons. (Boston Dynamics has them.)
Just a matter of time before they turn them on us.