Domingos: AI Alignment Community Unknowingly Building a Censor's Toolkit

pmddomingos · x · 2026-07-06

Pedro Domingos cited an ICML 2026 paper, pointing out that the AI alignment community is unwittingly building a "censor's toolkit." This has sparked criticism and ethical debates over whether alignment work could be repurposed for content moderation. It represents an ongoing debate in the value alignment space.

Original post →

More from AGI Musings

AGI Musings channel →