We should be talking about zombie reasoning

totally agree ... BUT:

When the first apes sorted stones, they didn't use any sorting algorithms and toddlers reason based on still limited/yet to develop and yet to be tutored/still untutored skills, knowledge and the systems that connect these two (CNS, ... bio-chemistry) ...

AIs are still just models and far away enough from intelligence. You could "script" (code algorithms to return) entire 24/7 news channels AND the news they broadcast and even the characters and the crews personal blogs and if you let some SSS graphics dudes work their magic only a few of those with an above 90% recognition rate (of GenAI content) would be able to spot "the slop", because it really wouldn't be slop anymore.

But it's not the AIs fault. So, if their parents put in the time and effort, the AIs could already fake being a grownup pretty damn well, as can over 90% of the human population, and that is, as we all know, enough to build all that is. But we don't need that kind of AI and those efforts and returns would serve no more than some peoples entertainment. There are more efficient and concise training methods, but we all know almost nobody will/ever really care(s) anyway ...

True AI won't see the light of day for a long time. While the concepts will serve all kinds of narratives pretty much forever, the distinction between "it's all algorithms!" and "how does the human brain do stuff?" isn't really worth it.

Zooming in on the parents and their reasoning, on the other hand, now that would be an interesting pursuit.