How do we square cognitive diminishment with powerful new abilities enabled by AI?

On one hand you unlock insane productivity gains and abilities to do things much faster.

On the other, you risk outsourcing critical thinking to controlled models.

The whole - if you don’t use it you lose it thing…

I see some people say, well, don’t rely on the answers you get from prompts. Ok. Fair enough. But what happens when even the non-prompted information becomes a product of prompting? At that point it’s impossible to tell what is original thought and what is a product of AI usage and influence. It feels inevitable that a bulk of knowledge will eventually be influenced by AI.

I’m not concerned about AGI influencing the whole of humanity because I don’t think AGI will be possible for hundreds of years yet, but there’s some concern for human influence/ coercion.

Do we even have a choice in how it all shapes up? Do open source models break the potential chains? Or do we all become skilled dummies over a long enough time frame?