> With the exception of Adversarial attacks on particular algorithms, no human is creating new Chess theory, discovering new openings, for example.

That's not true though? Human chessplayers are constantly exploring new lines, and engines are a useful tool but by no means the bottom line - indeed lines that were neglected because engines gave them low evaluations have been a fruitful source of ideas lately, because humans play differently from engines and may not be able to find the refutation.