A common thing I see is a baby animal needing rescue by a human (which it does) and it comes back later on and rewards the human with a gift of some kind it thinks is valuable.
I watch a few podcasts as well and there are more that have their scripts generated and voiced by AI
I saw a video I wanted to share with someone, but it was part of a compilation. So you just search for it, right?
So I searched "cat lets brick fall onto mouse" and got... 100000 AI generated videos of cats with bricks? And cats with mice and cats being rescued by people (like you said). But not the video I was looking for.
We've totally passed the point where real information is impossible to find anymore. Video generation was really out of reach / delayed for a long time, and honestly all of those probably have a digital watermark in them that could be detected. YouTube could have prevented this if they'd have just been more proactive with detection and filtration. A simple "AI generated" and "not AI generated" filter would have prevented this.
It will always be subject to the delay in detecting the bypasses of the latest AI techniques.
There's something with these compilations. Almost as if deliberately AI slop is mixed in to numb the public to it, or for some AI startup to testdrive on an unaware public how good their stuff is.
Take compilations of lightning strikes for instance. There's always a couple that are just too spectacular or just unbelievably. Like a ball lightning going across the street.
Another common theme is to kill rare animals to stage cool photographs. However, I don't assume AI slop will (fully) replace that cruelty, unfortunately. But maybe using AI slop will be easier than animal abuse/killing so that those business models run dry for the most part.
I've seen these, but I don't think the ones I've seen are AI generated (except sometimes for the video thumbnail). They tend have appropriate clips from the movie matched to what's being described, and I'd be surprised if an AI model could do that.
My guess is they're human generated attempt to profit off a movie while avoiding copyright enforcement from the movie's owners.
I've posted an article on this matter: https://news.ycombinator.com/item?id=46121555
Worst are imho on the regular long vids side, the geopolitical advisor deep fakes, giving background to the news. Some with well over a million followers. Many of those have the same "we are a fan of the real person" disclaimer, many have no disclaimer.
And no one in the comments, of which many look fake too, notices it is AI. That is the most scary part.
Many who notice won't bother commenting, because most who notice know how pointless that is (counterproductive in fact: a comment is an interaction, any interaction is a positive for the "content"). Those that do notice and comment are either drowned out by
• those too numbed on the brain to care, let alone notice, who lap it up, and praise it
• bots (either those being used to interact with the clip to drive it's interactions counters, or more general spam bots)
or if there is anyone/anybot monitoring the negative comments are removed.