Use of AI tools - brain rot, etc

My company like everyone else these days is 'all in on AI'. I am told about a hundred times a day that these are the future and I need to get on board or I'll be left behind. Being a person needing a job, I am in alignment with this idea. We have lots of very powerful CLI tools and as of 2026 they are loaded with context as well. I am impressed with a great deal about these tools, but I have a couple of observations: 1. they will waste lots of tokens and time based on false assumptions that I have corrected many times before 2. they love decompiling closed source code and suggesting fixes that violate license agreements 3. they often offer very complex solutions to simple problems 4. they fry my brain. this is the biggest one. if a problem is fairly complex, the ai provides a long justification and i will read it and try to understand exactly what it means. but sometimes after enough analysis of the text, i realize it is all gobbledegook and the ai is confused or hallucinating -- this process is absolutely exhausting to me. 5. 'arguing' with ai. if the ai tool makes an error or misunderstands the problem, i end up 'arguing' with it. this gives me a terrible weird empty feeling that i can't quite put into words. 6. my own skills are going to atrophy. the AI writes code in a different style then me, it takes me longer to read and understand it 7. The tools have this general unawareness of what they are doing. Sometimes the work is great, sometimes it is not, and if I give equal weight to every answer, i get exhausted. Sometimes it is like listening to a sage senior describe a complex problem, and other times it is like listening to an exasperated student try to describe a for loop. The problem is, the mouthpiece is the same Now, my questions for the group: 1. How do you put guardrails on your own usage of these tools? 2. Have you found efficient ways of working where you can use the good aspects of these tools, but ignore the bad aspects? 3. Generally -- how to cope?