People often compare AI to tools like calculators. Adopting those technologies has indeed affected how we learn and what we are capable of, but those are pretty intentional trade-offs, and even then we do put restrictions on how people use them. We still make children learn arithmetic, and when I was in school for geochemistry I had to derive equations or manually calculate things at least once before I used any automated tools. Once you get used to the tools, you honestly do lose the ability to easily do it the hard way, but I think the hope is that when you have done it the hard way, you keep an understanding of the process. You automate these parts when you hopefully have a good understanding of them that you can build on.
With LLMs, we see people using the tool to completely offload the thinking portion of the task. People I see using them are not doing the hard work, not even once. They are not struggling with the concepts. I think the core of the issue is that many people are under the impression that school is a thing you just need to get through, or that is done to advance a career. They are not looking at education and understanding as goals on their own merit. There are a lot of reasons for that, and I don't really blame most people who feel that way, but when you view school work as something you just need to get through without intrinsic value, of course you are going to shortcut it as much as is possible. Worse yet, I think you can make an argument that most people will not need math beyond a primary school level. We do all need critical thinking and analysis skills to survive in the modern world, and that is exactly what people are shortcutting now with LLMs.