Pretty easy to find LLM outputs that mirror how humans react to being, say, bullied. I find this entirely flimsy.