Bill Kristol: AI guardrails without enforcement, liability and penalties aren't real guardrails

Miles_Brundage · x · 2026-09-16

Bill Kristol argues that AI "guardrails" lacking teeth — enforcement mechanisms, liability, penalties, and the weight of government behind them — are not real guardrails. The take, reshared by former OpenAI policy researcher Miles Brundage, feeds into the ongoing debate over whether industry self-regulation can suffice without binding government oversight.

Original post →

More from AGI Musings

AGI Musings channel →