[rsynnott]

I mean, that seems like the obvious explanation. Most training material that LLMs have about 'AIs' (in the actual thinking machine sense) would be sci-fi, and it's an extremely common narrative.