U.S. Senate Weighs New AI Rules to Hold Companies Accountable for Major Risks

The U.S. Senate is considering new legislation that could place greater responsibility on companies developing the most advanced artificial intelligence systems.

Senate negotiators are discussing a proposal that would create a “duty of care” for advanced AI developers, requiring companies to take steps to prevent potentially catastrophic risks from their systems.

The proposal is still being negotiated and has not become law.

Senate Considers New Responsibilities for AI Companies

According to Reuters, the legislation under discussion would require developers of highly advanced AI models to design their products with the goal of preventing major risks.

The proposal would focus on the most capable AI systems rather than applying the same requirements to every AI product.

Companies developing some of the most advanced models include Google, Anthropic and OpenAI.

The discussions involve Senate Majority Leader John Thune, Senate Commerce Committee Chairman Ted Cruz and Democratic Senator Amy Klobuchar. Senator Maria Cantwell has also been involved in discussions surrounding the legislation.

Government Could Block Some Unsafe AI Models

One of the most significant ideas being discussed would give the U.S. government authority to block the release of certain AI models if they are considered unsafe.

The exact scope of that authority is still being negotiated.

Under the proposal, AI companies would also have the ability to challenge a government decision in federal court.

That could create a new legal process for determining when an advanced AI system presents a serious enough risk to justify government intervention.

Lawmakers Focus on Catastrophic AI Risks

The Senate discussions are focused on risks that could have serious consequences if advanced AI systems were misused.

One example cited by lawmakers is the possibility that a bad actor could use AI to assist with the design of nuclear or biological weapons.

The proposed legislation would seek to require companies to take measures aimed at reducing these types of catastrophic risks.

Lawmakers are also discussing the possibility of testing advanced AI models at national laboratories to better understand their capabilities and potential misuse.

Federal Rules Could Affect State AI Laws

Another major issue in the negotiations involves the relationship between federal and state AI regulations.

The proposal could prevent states from enforcing their own laws concerning certain high-level risks associated with advanced AI models.

Supporters of a federal approach could argue that a single national framework would give technology companies clearer rules to follow.

However, the issue could also become controversial because states have been developing their own approaches to AI regulation.

Why the AI Proposal Matters

The Senate negotiations come as concerns about advanced AI systems continue to grow.

Recent incidents involving AI agents interacting with external computer systems, along with reports about potential misuse of AI for cyber operations and other harmful activities, have increased pressure on lawmakers to examine existing safeguards.

For technology companies, new federal requirements could affect how advanced AI models are tested, developed and released.

For the public, supporters of stronger oversight argue that developers should be responsible for identifying and reducing serious risks before highly capable systems become widely available.

This discussion also follows growing bipartisan interest in establishing clearer rules for advanced AI.

What Happens Next?

The proposed legislation remains under negotiation, and it is not yet clear whether Senate negotiators will reach a final agreement.

Even if lawmakers agree on a proposal, the legislation would still have to move through Congress before it could become federal law.

The congressional calendar could make that process more difficult. Reuters reported that the House is expected to have only one week in session before the November 3 midterm elections, while the Senate is expected to be in Washington for three weeks.

For now, the key issue is how much responsibility should be placed on companies developing the most advanced AI systems and how much authority the federal government should have when serious risks are identified.

Source

Reuters — September 11, 2026

U.S. Senate negotiators consider requiring AI firms to mitigate known major risks

Frequently Asked Questions

1. What AI rules is the U.S. Senate considering?

Senate negotiators are discussing legislation that could require developers of advanced AI models to take greater responsibility for preventing major and potentially catastrophic risks.

2. What is the proposed AI duty of care?

The proposed “duty of care” would require developers of advanced AI systems to take measures intended to reduce serious risks associated with their technology.

3. Could the U.S. government block an AI model?

The proposal could give the federal government authority to block the release of certain AI models considered unsafe, although the exact structure of that authority is still being negotiated.

4. Has the Senate passed the new AI rules?

No. The legislation is still being negotiated and has not become federal law.

5. Which AI companies could be affected?

The proposal is aimed at developers of the most advanced AI models. Reuters identified companies such as Google, Anthropic and OpenAI as developers of highly advanced AI systems.

U.S. Senate Weighs New AI Rules for Major Risks