so.. heres a reason to both be less scared of AI, and more scared of AI:

stuff like google using gathered data to recommend you stuff that you may like, like with youtube recommended videos, because it "knows" all your preferences.

mindblowing fact for you: AI has no idea. the only reason why AI recommendations would work perfectly for you, is because you have very mainstream interests, because it can only base itself off of links it has already connected. (as in, someone else who watched the same videos as you, also watched this one, so therefore you will like it too) since the links you make may not be the same as those someone else makes, this is incredibly flawed. and ofcourse the AI has no idea of *why* you may have watched a video.

protip if you want to REALLY upset youtube algorithm..

watch this video;

it is a long format video of a very specific interest, you may watch it with very different intent, but the algorithm doesnt know that, so it assumes you must REALLY like cheese. the algorithm doesnt understand human things like watching something with other intent, like watching thunderdome '98 to laugh at the people, so as a result will assume you like watching this stuff in general.

as a result, there's less reason to be scared of AI, because in truth it is a completely unaware formula trying to approximate by comparing you to others like you.

but as a result, there's more reason to be scared, because if we then start to see the output of these tools as a "truth", it generates more false positives than it does perfect matches. it cannot distinguish a possible terrorist from a science professor looking into fun experiments for the classroom, and when it starts to assume traits based on these "false matches" you end with a ban of ice cream because statistics show that it causes people to drown in swimming pools.