> One major problem I see with the use of AI is that it will prevent people from building an understanding of <insert problem domain X here>. This will reduce people's ability to drive AI correctly, creating a circular problem.

Very much the opposite. LLMs do a fantastic job of increasing accessibility of knowledge.

They have wide exposure to content and is incredibly good at spotting patterns and suggesting both well established norms from the current domain and serendipitous cross domain concepts.

I feel the concern you share is LLMs expose new frontiers to people who otherwise might not even have imagined that frontier exists and then those people do a lazy or superficial job of it because they lack any internal motivation to do a deep dive on it