AI researchers debate the blurry line between AI ethics and AI safety communities

Dr_Atoosa · x · 2026-11-21

A discussion on the ethics/safety divide in AI: the reply argues these labels are imperfect sociological distinctions — tools for understanding historically formed, overlapping subcultures, their practices, networks, power relations and visions — not community identities, with boundaries shifting over time.

@dioscuri admits he can't tell which side his own work falls on, has worked with brilliant ethicists at CFI and beyond, and only laments that some of the loudest recent voices were dogmatic about the tech. The reply notes specific dogmatic voices don't represent the wider communities.

Related event: AI researchers question the ethics/safety community divide(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →