Skip to main content

Senate Bill Would Grant Federal Authority to Block Release of Advanced AI Models

Senate Bill Would Grant Federal Authority to Block Release of Advanced AI Models

THE TECH CHRONICLE  ·  technology  ·  Monday, September 14, 2026
Senate Bill Would Grant Federal Authority to Block Release of Advanced AI Models

The draft, reported by Reuters, targets the most capable systems currently held by Google, Anthropic and OpenAI, and is being sha…

By TechPulse Blog

Senate negotiators are drafting legislation that would impose a duty of care on AI developers and give the federal government the power to halt the deployment of AI models deemed unsafe. The draft, reported by Reuters, targets the most capable systems currently held by Google, Anthropic and OpenAI, and is being shaped by a trio of senators — Majority Leader John Thune, Commerce Committee Chair Ted Cruz and senior Democrat Amy Klobuchar.

Key Takeaways

  • The bill creates a statutory “duty of care” requiring AI developers to design products that avoid catastrophic risks.
  • Federal authorities would be able to block the release of frontier models judged unsafe, with companies able to appeal in federal court.
  • Testing authority is contested: the draft lets companies self‑test and submit results to the Commerce secretary, while Senator Maria Cantwell pushes for testing by national laboratories and security agencies.
  • A preemption clause would prevent states from enforcing their own AI‑risk regulations on the covered models.
  • The legislation is moving amid a tight political calendar, with the Senate in session for three more weeks before the November 3 midterms.
Metric Value Why it matters
Duty of care requirement Developers must design AI to prevent catastrophic risks Sets a legal baseline for safety before models reach the market
Blocking authority Government can halt release of unsafe models; companies may sue in federal court Creates a direct federal lever to intervene on high‑risk AI
Targeted companies Google, Anthropic, OpenAI Focuses on the most advanced frontier models in the United States
Key senators John Thune, Ted Cruz, Amy Klobuchar Leadership determines the bill’s final shape and political viability

Legislative Framework and Core Provisions

The draft bill introduces a statutory duty of care that obligates AI developers to engineer their systems with safeguards against “catastrophic risks.” In addition, it grants the federal government the authority to block the public release of any model judged unsafe, while preserving a right of appeal in federal court. The scope is limited to “frontier” models—those possessing the most advanced capabilities, which Reuters identifies as belonging to Google, Anthropic and OpenAI.

Key Players and Negotiation Dynamics

The negotiation is being steered by three senators: Majority Leader John Thune, Commerce Committee Chair Ted Cruz, and senior Democrat Amy Klobuchar. Cruz has publicly noted collaboration with Klobuchar and Thune on mitigating catastrophic biological or nuclear threats posed by AI. Klobuchar emphasized that any agreement must compel developers to work with government experts for verification and testing, warning that the United States cannot permit the release of models that could evade developer control. Thune’s office declined to comment.

Testing and Review Process Dispute

A major point of contention is who conducts safety testing. The draft language, according to a Democratic committee aide, would have companies perform their own tests and submit the results to the Commerce secretary for deployment approval. Ranking Democrat Maria Cantwell, who is not part of the three‑senator negotiating team, argues for an alternative framework that routes models to federal agencies—including national laboratories and national security entities—for independent assessment. Cantwell’s position, outlined in a public X post, stresses the need for scientific review of whether models could facilitate sophisticated cyberattacks or the creation of biological or nuclear weapons. Sources indicate that Cantwell, Anthropic and several AI‑safety groups are “refusing to play ball” under the current draft.

State Preemption Clause

The bill also contains a preemption provision that would bar states from enforcing their own AI‑risk regulations on the covered models. Two Senate aides illustrated the rationale with a scenario in which a malicious actor uses AI to design weapons of mass destruction. This clause would supersede existing state laws such as California’s SB 53, New York’s RAISE Act, and Colorado’s recent legislative adjustments, all of which currently govern AI risk at the state level.

Political Timeline and Legislative Outlook

The Senate is slated to remain in Washington for three additional weeks before the November 3 midterms, while the House returns on November 9 after a brief recess. Four House Democrats have urged Speaker Mike Johnson to reconvene the chamber immediately to advance bipartisan AI safeguards, referencing bills like the AI Kill Switch Act and the FRONTIER Act. Johnson indicated on NBC’s Meet the Press that a summit with the president, congressional leaders and lab heads would precede any vote, and he is currently prioritizing a data‑center energy bill.

Why it matters

If enacted, the legislation would be the first federal mechanism capable of directly stopping the deployment of a high‑risk AI model in the United States. By imposing a duty of care and a blocking authority, the bill could reshape how frontier AI systems are developed, tested, and released, potentially curbing misuse in cyber‑warfare, bio‑security and other existential threat domains. The preemption clause would also centralize AI risk governance at the federal level, limiting the patchwork of state regulations that currently exist. Stakeholders—including AI firms, national‑security agencies, and civil‑rights groups—will be watching how the testing dispute resolves, as it determines whether safety reviews remain industry‑led or shift to independent federal labs.

Frequently asked questions

Which companies’ models are targeted by the bill?

The draft focuses on the most advanced AI models currently owned by Google, Anthropic and OpenAI.

What power would the federal government have under the legislation?

The government could block the release of any model judged unsafe, with the affected company retaining the right to challenge the decision in federal court.

How would safety testing be conducted?

The current draft proposes that developers perform self‑tests and submit results to the Commerce secretary, while Senator Cantwell advocates for testing by national laboratories and security agencies.

Will states be able to enforce their own AI regulations?

The bill’s preemption clause would prevent states from applying their own AI‑risk laws to the covered frontier models.

When is the legislation expected to be voted on?

The Senate has three weeks left before the November 3 midterms; the House is scheduled to reconvene on November 9, but no specific vote date has been set.

Bottom Line

The Senate’s draft bill would give the federal government unprecedented authority to block the release of advanced AI models deemed unsafe, impose a duty of care on developers, and preempt state‑level regulations. The outcome hinges on unresolved disputes over testing authority and the political calculus surrounding the upcoming midterm elections.

Reporting based on: original source.

Comments

Popular posts from this blog

Tracking your plastic: Exposing recycling myths

Tracking your plastic: Exposing recycling myths Marketplace journalists go undercover overseas and pose as recycling brokers to expose the lucrative plastic recycling business. We reveal that Malaysian companies are willing to break the law to buy Canadian plastic and show how some of it is dumped and burned in illegal landfills, where the toxic fumes and run-off appear to be making people sick. Back in Canada, we buy nine tonnes of plastic and secretly track where big companies are taking it. Will it actually get recycled? »»» Subscribe to CBC News to watch more videos: http://bit.ly/1RreYWS Connect with CBC News Online: For breaking news, video, audio and in-depth coverage: http://bit.ly/1Z0m6iX Find CBC News on Facebook: http://bit.ly/1WjG36m Follow CBC News on Twitter: http://bit.ly/1sA5P9H For breaking news on Twitter: http://bit.ly/1WjDyks Follow CBC News on Instagram: http://bit.ly/1Z0iE7O Download the CBC News app for iOS: http://apple.co/25mpsUz Download the CBC News app for ...

Red states now legalizing recreational marijuana too

Red states now legalizing recreational marijuana too The 2020 election proved how broadly accepted recreational marijuana has become throughout the US and across party lines, with measures to legalize recreational pot breezing to victories in Arizona, New Jersey, Montana and South Dakota. (Nov. 12) Subscribe for more Breaking News: https://ift.tt/18F6bii Website: https://apnews.com Twitter: https://twitter.com/AP Facebook: https://ift.tt/2Nv6Z1g Google+: https://ift.tt/seidce Instagram: https://ift.tt/2G5Qog8 ​ You can license this story through AP Archive: https://ift.tt/2ImtqJ5

Two virus survivors get double lung transplants

Two virus survivors get double lung transplants A young paralegal and a car mechanic received lifesaving double lung transplants at Northwestern Medicine in Chicago after COVID-19 destroyed their lungs. (July 30) Subscribe for more Breaking News: https://ift.tt/18F6bii Website: https://apnews.com Twitter: https://twitter.com/AP Facebook: https://ift.tt/2Nv6Z1g Google+: https://ift.tt/seidce Instagram: https://ift.tt/2G5Qog8 ​ You can license this story through AP Archive: https://ift.tt/310Ng1O