When you are outspokenly and practically anti-AI, it may sometimes appear that you are rejecting all code written (or generated) by others, and insisting on writing everything yourself. That is an easy trap to fall into, especially if you are already prone to "not invented here" syndrome — as many Linux users and people in the greater opoen source community sometimes are. It is important not to, because no one person can recreate the entire digital ecosystem themselves.
But rejecting AI does not ultimately come from the same place as "not invented here" syndrome. Rejecting AI still means using libraries and existing services or applications written by skilled and experienced humans. Rejecting AI still means drawing on other human intelligences and the sum total of accumulated human knowledge. Rejecting AI still means building software as part of a team with relationships to other teams in the "outside world" — only with a few additional requirements on who the other teams and other team members are, and how they do their jobs. Rejecting AI is a filter for quality, not a rejection of everything other than what you do yourself.
If you like, it may also be a political stance. But that stance is not (or should not be) hardcore, total libertarianism, where you have to be absolutely self-reliant and cannot count on anyone else's code. That philosophy leads to a whole slew of nasty outcomes when it collides with reality — and if you get nasty outcomes in software, you get them in real life. Instead, the strongest and most viable anti-AI political stance is rooted in respect and recognition of the essential human element in all creative endeavors, including software engineering. Under that worldview, there is no intrinsic need to do everything yourself unless you gain something from it, but it is important to know where the code you use comes from, who wrote it, and who you can ask questions about it and report bugs to.
Do things yourself if you can do them better than other people, or know that other people will not be able to understand your need well enough to build something that will solve your problems. Do things yourself if you enjoy them, gain useful experience or knowledge from doing them, or otherwise benefit from doing the work yourself. Avoid, on the other hand, doing things just because you are unwilling to use code that someone else wrote.
But of course, do not use AI. Use real libraries written by real people.