If you watch a grand master at chess, and watch what the master did to win, you'd learn some options.

If you do that with lots of chess experts, you'd learn a lot of options. You'd learn a lot of strategies.

If a few of those are AI, you're still only going to learn options and see strategies.

I'm not sure you've played chess if you think you can just memorise a pattern and play the pattern. No matter how many AI games you've watched or studied, every time you play a person, you've got to select from what you know to decide what to do.

The cheating part is a different issue.

But the fear that a bunch of players will learn from AI means nothing. Let's say five years from now the two top players both learned by studying AI played games. They still end up being two humans playing against each other, not two robots.

And the reason is in the start of your article. An AI can look way into the future of all possible moves stemming from the current situation and current move. A human is limited as to how far they can look ahead.

This means that even if a human does the move an AI might have done, they can't then do the analysis to decide on what to do after that.

Humans are not AI. No matter who they learn from, they'll still play like a human.