The paradox of AI automation at the workplace

I find it interesting how people are very ambitious to automate their processes at their companies using new AI tools, but they set the limit of automation on their own terms, and if you try to pass it, you become the cynic or negative one.

For example, this guy where I work suggested to use LLMs before 1:1 meetings to scan all the data on them from Google Drive and Slack conversations, so they can be better prepared for the conversation. First this sounds very useful, but if you think about it, it tries to automate a process where humans are involved, and just makes things even more detached and robotic, in a world where we are already obliged to minimize human interactions.

But then, lets say you like their idea, and even give feedback: Oh, so what if maybe you just create a video chatbot that can just join the meetings instead of you, and use all that gathered data to talk to them? Best case, all of a sudden it becomes "too much", and make them feel like you overstepped, as they realize how they are on a path to automate themselves out of their jobs. Worst case, they agree and start working on implementing it...

So in conclusion, who defines where the line is to call something a good business improvement and actively contribute to it, and why do so many people not realize that they are chopping the wood below themselves?