Models respond humanly because they are trained on our data

TheStalwart · x · 2026-08-31

Discussion on whether AI models should be anthropomorphized. The argument suggests that since models are trained on human data and fine-tuned to react in ways we find desirable, responding in human-like ways is a natural outcome. It also notes that using anthropomorphic language for pragmatic purposes often carries implicit assumptions that should be made explicit.

Related event: Debate Over AI Anthropomorphism: Fear vs. Accuracy(4 posts)→

Original post →

More from AGI Musings

AGI Musings channel →