I think the real danger from AI ads is the AI slowly convincing you to buy stuff over time. It's going to be super effective with the less technically adept.
(man bash: ${X:-abcde} means use value of X else abcde by default )
where GOAL can be anything one may be persuaded into. Choosing $15 burger over $7, brand X over brand Y, notion of X wholesale (to EV or not to EV), Elections, climate-this-or-that, ...
People are massively using it as search engine. So it does not need to lie, just can "spare"/not-show some results that do not match the GOAL..
extra: like 25th frame in 24frame video.. like, last word at each sentence is part of another 5-word goal-sentence.. repeated across 20 sentences..
IMO the "text exegesis" (i.e. what particular text actually means) may need resurrecting as discipline, and not only in higher-education / academia but down into school.
scraping all history every time may or may not be possible..
should one have like 5 accounts and share them with 10 people? across the globe?
mmh. interesting times ahead
I've found that more often than not, it gets at least one key feature/option/etc. outright wrong whenever I've tried that, making it effectively useless for me. Since I need to verify the exact information myself anyways, I'm 90% the easy to just having the different items in comparing up in side-by-side browser tabs, anyways.