Sep 18, 2026 · 35 min
Lina Khan challenges AI’s catastrophe-first regulation
Lina Khan on the AI-Safety Panic
The episode tests whether existing legal powers can address AI’s immediate economic harms while companies shape the rules meant to govern them.
- 1Existing civil and criminal laws may reach harmful AI behavior, but opacity complicates accountability and enforcement.
- 2Khan shifts attention from hypothetical extinction scenarios toward workers, affordability, market concentration, and corporate power.
- 3Public agencies still possess investigative and oversight tools, but political influence and information gaps threaten effective regulation.
Don't miss
Lina Khan argues that governments and Congress have more oversight and information-gathering power than AI companies suggest, if institutions use it.
The brief
Charlie Warzel and Lina Khan begin with the familiar warnings: increasingly autonomous AI could become uncontrollable or catastrophic. Khan asks whether those warnings obscure a more immediate regulatory question.
The conversation tests existing civil and criminal law against opaque systems, unexpected behavior, and unauthorized activity, asking what companies knew and what they should have done to prevent harm.
Khan redirects the debate toward workers, affordability, market concentration, and corporate power, arguing that regulators should investigate unfair practices rather than wait for speculative disasters.
The episode also examines regulatory capture: AI companies and political allies may define the terms of oversight, while executives’ financial incentives complicate their public safety warnings.
Khan’s clearest conclusion is institutional rather than technological: governments and Congress retain oversight tools, but must close the public’s information disadvantage before companies set the rules.