What Is Congress Doing About AI? Moratoriums, Guardrails & the Oversight Debate

If you’ve followed the news since the November 2026 midterms, you’ve probably heard that Congress wants to “do something” about AI. But what exactly? The proposals on the table range from pausing advanced AI development entirely to targeted safety rules and new oversight tools, and the debate over which approach works best, including whether a pause on advanced AI can be enforced, is just getting started.

Why AI Is Garnering Interest

AI moved from tech pages to the ballot box this year. An NBC News exit poll found that 56% of voters called AI and data centers a crucial or major factor in their vote for Congress. With Democrats taking control of both chambers in January, lawmakers have signaled that AI legislation will feature prominently in the new session.

Approach 1: A Moratorium on Advanced AI

The most sweeping proposal is the Ban Artificial Superintelligence Act of 2026, introduced in September by Sen. Bernie Sanders (S. 5493) and Rep. Greg Casar (H.R. 10538). The bill would create a cabinet-level Department of Artificial Intelligence and pause development of advanced AI systems trained above a threshold of 10^25 operations until that department sets safety rules. It would also ban artificial superintelligence outright.

Supporters argue that the technology is advancing faster than safeguards, and that a pause would give regulators time to catch up. The bill is currently in committee.

Approach 2: Targeted Guardrails and Oversight

Other lawmakers favor narrower, enforceable rules. Proposals discussed in Congress include safety testing and auditing requirements, whistleblower protections, “kill switch” mechanisms that would let systems be shut down, and incident reporting. House Democratic committee leaders have already asked several frontier AI companies to report safety incidents to Congress within 48 hours of detection.

This is where much of the technical debate sits. Critics of a pause argue it would mostly affect developers who follow the rules, while capabilities continue to spread through open-weight models and overseas development. Fran Horvath, Chief Vision Officer at AI Standards, makes this case in a recent report, proposing verifiable technical controls as an alternative, such as cryptographic identity for AI systems, continuous runtime monitoring, and tamper-evident audit trails that regulators can check.

Approach 3: Who Writes the Rules, Washington or the States?

A parallel fight is over federal preemption. States including California, New York, and Illinois have passed laws targeting frontier AI models. The White House’s National Policy Framework for AI, released in March 2026, urged Congress to preempt state AI laws it considers burdensome. Opponents, including many state lawmakers, argue that states should stay free to set stronger protections. Preemption efforts stalled in 2026 and appear less likely under the new majority.

What Happens Next?

Even with new majorities, comprehensive AI legislation faces real hurdles. Most bills will need 60 votes in the Senate, and the administration has opposed binding AI safety mandates. Many analysts expect hearings, oversight, and targeted measures, such as kids’ online safety rules, to move faster than sweeping changes like a moratorium.

What It Means for Businesses

For now, state laws remain in effect, so companies using AI should keep complying. It’s also a good time to document how AI systems are used, monitored, and approved. Whatever form federal rules take, organizations that can show clear records of their AI activity will likely be better prepared.

The Takeaway

Congress is now debating how to regulate AI: pause it, set guardrails, or require proof that systems behave as intended. The answer will shape how AI is built and used for years to come. For a deeper look at one side of the debate, this analysis of the limits of an AI pause explains the case for cryptographic controls in more detail.

AI Standards Inc
City: Lanett
Address: 311 13th Ave South
Website: https://aistandards.inc
Phone: +1 678 205 9720
Email: press@aistandards.inc