If all we were seeing is the prompt used to generate the video, then there wouldn’t be a problem. Human-written fiction is generally valuable.
Instead we’re getting single sentences masquerading as “a picture worth a 1000 words”, or worse with video. Only 1% of that is actually the valuable part (the prompt), the rest is filler words and hallucinated slop.
A video or picture of reality is inherently more informative than any AI generation.