OpenAI trains on our aggregated thoughts—attribution is the real problem

arjunrajlab · x · 2026-09-09

A researcher raises a thought-provoking point: OpenAI and peers are training on our publicly shared ideas in aggregate, and dissemination is arguably good for science. The real problem is attribution—since humans are, depressingly, self-interested, creators get no credit once their ideas are absorbed into an anonymous model. He even floats the reverse idea: maybe non-anonymity would be better.

Related event: Navier-Stokes dispute raises questions of attribution for aggregated ideas(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →