Using ChatGPT doesn't dumb down your students. Not knowing how it works and where to use it does. Don't do silly textbook challenges for exams anymore - reestablish a culture of scientific innovation!
LLMs can't produce intellectual rigour. They get fine details wrong every time. So indeed using ChatGPT for doing your reasoning for you produces inferior results. By normalising non-rigorous yet correct sounding answers, we drive down the expectations.
To take a concrete example, if you tell a student to implement memcpy with chatgpt, and it will just give an answer which uses uint64 copying. The student has not thought from first principles (copy byte by byte? Improve performance? How to handle alignment?). This lack of insight in return to immediate gratification will bite later.
It's maybe not problem for non-STEM fields where this kind of rigor and insight is not required to excel. But in STEM fields, we write programs and prove theorems for insight. And that insight and the process of obtaining it is gone with AI.
You need to know how to do things so you know when the AI is lying to you.
High schools are a long way off from that level of education. I took AP CS in highschool and it was a joke by comparison. Of course YMMV. The best highschool CS course might be better than the worst university level offerings. We would always have know it all students who learned Java in high school. They either appreciated the new perspective on the fundamentals and did well, or they blew off the class and failed when it got harder.
I also worked for the faculty for the better part of my university studies, and I know that ultimately changing the status quo is most likely impractical. There are not enough resources to continuously grade open-ended assignments for so many students and they probably need the pedagogical pressure to learn fundamentals. Still makes me a bit bitter from time to time.