I think this may be just interpolation weirdness. Apparently the video was recorded in 24fps but encoded in 60.
See this analysis, I posted it in another comment: https://youtube.com/shorts/hXhDB1fK7ok
Edit: worth noting the interpolarion may be done by a similar AI engine which explains why it looks so similar to AI generated footage. Similar technology as the motion smoothness on TVs or DLSS frame gen. If it truly is 24-60fps and the interpolation is done by AI this would indicate over half the frames are AI generated. Would explain a lot while not necessarily meaning the video as a whole is AI.