Comment by haldujai

No, that’s not my argument. Rather it is that a counterexample to a mathematical proof whether produced by human or AI may be a scientific discovery but it is not inherently new and brilliant by definition.

The case for AI is also weakened when the model is steered by an expert.