Artificial Intelligence ≠ Thinking

I have been reading a book called The Future by Naomi Alderman, that has an interesting discussion around what Artificial Intelligence is and is not. A lot of people not versed in the nuances of the technology think it means computers are evolving and gaining the ability to actually think. 

One of the characters in the book is a college instructor who is trying to explain to his students what AI is. He uses an example of a computer learning how to play tic-tac-toe. The computer can be programmed to determine all of the (362,880) iterations that can be made in the game, within seconds. The problem is, the computer also needs to be programmed to have an indicator of what is a win and what is not. Once that information is provided, the computer can win or tie every game against an opponent. That doesn't mean its thinking. All it means is, the computer can sort through the possible moves, and the potential outcomes, and determine the best option according to its programming.

AI functionality like ChatGPT is a much more powerful implementation of the same mechanisms, but essentially the same process.

About twenty-five years ago, I wrote a couple of programs for the GEOS operating system on Commodore 64 that are still available for download. One was Tic-Tac-Toe and the other was a variation of the game NIM. In both cases, I had to develop the logic around winning the game, and then code that logic in Assembler (shudder), to make it work. There was no AI involved, since I defined the algorithm and logic for the game's strategy.

But with the advancements we have seen in computing power over the past two decades, the systems can now process information, recognize patterns and make decisions based on that data in a fraction of the time that was necessary before. But those systems don’t understand or consciously think. They are programmed to do those tasks. 

A metaphor that ChatGPT provide to me when I was writing this was:

AI is like a really smart calculator that knows how to guess based on huge amounts of information — not like a brain that’s aware it’s guessing.

Does that mean that computers won't be able to think in the future? Hard to say. Thinking involves using your mind to understand, reason, decide, imagine, plan, or reflect. But how our minds actually perform those tasks, we don't really understand.

AI can simulate some results of thinking (like solving math problems or generating text), but it doesn’t have consciousness, awareness, or intentions. It doesn’t know anything in the way you do — it’s just processing inputs and generating outputs based on training data.