Sens. Josh Hawley and Richard Blumenthal introduce a bill denying Section 230 protections for generative AI content; experts are split on whether §230 applies
Sens. Josh Hawley and Richard Blumenthal want to clarify that the internet's bedrock liability law does not apply to generative AI …
Context & Ripple Effects
This is Sen. Josh Hawley's third run at rewriting Section 230, but his first aimed at a technology rather than platform behavior. His earlier efforts — a 2019 bill stripping protections from large platforms and a Trump-administration-backed effort in 2020 that reviewers called largely toothless — went nowhere. Pairing with Democrat Richard Blumenthal changes the calculus: the two already collaborate on tech accountability, from the bipartisan Senate bill making platforms liable for their recommendation algorithms to KOSA.
The bill also lands in a live doctrinal fight. Legal experts remain divided on whether Section 230 ever covered machine-generated content, with several arguing after Justice Gorsuch's 2023 comments that the law will not shield AI firms from lawsuits over outputs. If the skeptics are right, Hawley and Blumenthal are legislating a clarification of an exemption that may not exist — if the optimists are right, they are closing a genuine loophole.
First-order effects
- Generative AI developers lose the ability to lean on the internet's intermediary liability shield for their outputs, meaning defamation or harm claims over generated content would attach to the model provider rather than dissolving under Section 230.
- Because experts are split on whether the shield applies to generated content at all, the bill's immediate effect is to force that unsettled question onto the congressional and courtroom agenda before litigants resolve it case by case.
Second-order effects
- Platforms embedding generative features would respond by demanding indemnification clauses from model providers, pushing liability cost down the AI supply chain toward whoever trains the model.
- The bipartisan pairing separates this bill from Hawley's earlier solo attempts, giving Section 230 opponents a template: amend the statute one technology at a time, as the later recommender-algorithm amendment did.
Third-order effects
- If Congress keeps exempting specific technologies — first recommendation algorithms, now generative models — Section 230 erodes from a uniform distributor shield into a sector-by-sector liability map, with each new AI capability triggering its own carve-out fight.
- AI firms would internalize legal review of outputs as a core product function rather than a compliance afterthought, mirroring the shift content moderation underwent once platforms lost confidence in blanket protection.
The trend: Section 230 is being carved up category by category as bipartisan lawmakers extend liability from distributors of content to the systems that generate it.