The opportunity for Congress to pass significant artificial-intelligence legislation before the midterm elections is effectively closing, though a few senators continue to pursue more limited measures.
What proposals are in play?
- Sen. John Kennedy (R-La.) plans to seek unanimous consent on Wednesday for a bill that would require companies, rather than the government, to control an AI "kill switch." Kennedy told Semafor he is not sure whether anyone will object but believes this approach is the most feasible in the near term.
- Sens. Josh Hawley (R-Mo.) and Richard Blumenthal (D-Conn.) are pushing for a floor vote on their AI safety proposal, attempting to bring their measure before the full Senate.
OpenAI's perspective
Chris Lehane, OpenAI's chief global affairs officer, has advocated a three-pronged regulatory strategy: strengthening industry safety practices, enacting new legislation, and pursuing international leadership on safeguards under the Trump administration. Lehane also agreed that significant legislative action before Election Day is unlikely.
Why this matters
Because substantial congressional action appears unlikely before the midterms, near-term management of AI risks will rely more on industry practices, executive-branch guidance, and any limited statutes that can be advanced quickly. Senators noted there might be more opportunity during the post-election "lame duck" session, but broader AI bills are seen as difficult to pass in the immediate timeframe.
Next steps
Progress depends on whether senators can secure unanimous consent or force a floor vote. Until then, tech companies' internal safety measures and international coordination efforts are likely to shape how AI risks are addressed through the upcoming election and afterwards.
