Congress Has Introduced Dozens Of AI Bills. Here’s What They Would Do.

Lawmakers introduced dozens of bills to address concerns about artificial intelligence (AI) as tech leaders issued dire warnings about its threat to humanity.

Tech leaders and AI researchers warned that the technology’s development must slow down, as Jacob Coxon, a former researcher at Anthropic, said people working at AI companies were “genuinely frightened” about the “fate of humanity in the next two years.” Several bills that would impose more federal regulations on AI remained untouched as these warnings garnered worldwide attention.

Independent Vermont Sen. Bernie Sanders and Democratic Texas Rep. Greg Casar introduced the “Ban Artificial Superintelligence Act,” which would ban AI superintelligence that could “surpass human intelligence or have the capacity to overthrow human governments” and impose a pause on advanced AI development until a cabinet-level federal agency is established to “safeguard the public from the dangers” of AI.

An individual who created such technology could face up to 20 years in prison and a company that violated the ban would face a corporate death penalty, meaning the business would immediately be shut down.

Sanders introduced another bill called the Artificial Intelligence Data Center Moratorium Act in March, which would prohibit the construction or upgrade of AI data centers until comprehensive federal safety laws and worker protections were enacted. It restricts the export of AI-related computing hardware to countries without similar laws in effect.

Republican California Rep.  Jay Obernolte and Democratic Massachusetts Rep. Lori Trahan introduced the Frontier Risk Oversight, National Transparency, Independent Evaluation, and Reporting (FRONTIER) Act in July. The legislation intends to provide federal oversight of the “development and deployment of frontier artificial intelligence.” The bill focuses exclusively on major developers creating the most powerful “frontier” AI models rather than smaller innovators and startups.

The legislation would create a licensed marketplace of independent verification organizations (IVOs) to conduct external evaluations and assess safety testing. It would require major developers to report critical safety incidents to the federal government.

The bill would empower the secretary of Commerce to issue emergency orders to suspend the development or deployment of AI models if they pose an imminent threat. It aims to set a single federal baseline that preempts state-level laws on third party auditing, risk disclosures and incident reporting.

Olbernote introduced The Creating Resources for Every American to Experiment with Artificial Intelligence (Create AI) Act to establish the National Artificial Intelligence Research Resource NAIRR), a national program to provide researchers, educators and students with access to AI data, computational resources and educational tools. The program would be established by the National Science Foundation (NSF).

The legislation would require NSF to select a nongovernmental organization to operate NAIRR to ensure that an annual allotment of computational resources is provided to projects focused on AI privacy, safety and security.

The bill stated that access to computational resources and datasets necessary for AI research is often limited to large technology companies and that more researchers were needed to ensure that AI “is developed in a manner that benefits all people of the United States.”

A bipartisan bill called The Resources for Evaluating and Documenting (READ) AI Models Act would direct the National Institute of Standards and Technology (NIST) to develop a template for information to accompany AI models. The template would be used to document information about an AI model and any associated data, such as the model’s name, its developer and the cutoff date for the model’s training data.

The template would include the name of the model, the identification and the location of the developer, the release date and the languages supported by the model.

The director of NIST would have to submit an assessment of the program’s effectiveness to Congress 12 months after establishing it. If the program were effective, the director would provide a permanent implementation and make a publicly available website providing information from the template.

Another bill called the Protecting Consumers from Deceptive AI Act would require NIST to establish task forces to develop standards and guidelines so individuals could identify generative AI content. The technical standards would provide watermarking and digital fingerprinting for audio and visual content.

The House passed the Small Business Administration Artificial Intelligence Utilization Act of 2026 in June, which would amend the Small Business Act to require the Administrator of the Small Business Administration to report on the use of AI and machine learning by the administration. It currently sits in the Senate Small Business and Entrepreneurship Committee.

Republican New York Rep. Mike Lawler and Democratic New Jersey Rep. Josh Gottheimer put forth the Stop Rogue AI Act Tuesday, which aims to ensure businesses and federal agencies could see “rogue and dangerous AI agents operating inside their own networks.” The legislation would direct NIST to develop national standards, guidelines and practices for “discovering, verifying, and controlling AI agents.”

The bill does not make private companies legally obligated to follow the standards.

Republican Texas Rep. Nathaniel Moran introduced the AI Incident Reporting Act in June, which would require certain AI developers to submit reports to the secretary of Commerce about any concerning activity, including behavior suggesting that the model is attempting to evade human oversight, resist shutdown or obtain unauthorized access to tools, systems or privileges.

The legislation would require developers to report any evidence that a model demonstrated the ability to accelerate or automate research, development, evaluation, engineering or improvement of advanced AI systems. Any incident of harm or possible risk to U.S. national security would need to be reported to the Commerce secretary.

The Open-Source AI Leadership Act would direct the Commerce secretary to support the adoption and use of American open AI models, although the secretary would not be required to support it. An open-source model is a system whose architecture, training code and components are freely shared to any who can use, modify or distribute them.

The legislation acts as an assessment and reporting directive rather than an enforcement mechanism with legal penalties. It would direct the Commerce secretary to review any relevant program related to the adoption and use of open models and enter into agreements to support its adoption.

Lawmakers were not keen on supporting a ban on AI superintelligence in discussions with the Daily Caller News Foundation Thursday.

“No, I’m not going to support that right now,” Republican Louisiana Sen. John Kennedy told the DCNF. “I’ll support a light touch system of rules that we all agree to, and hopefully China does too, and that looks like somebody designed them on purpose.”

Republican Missouri Sen. Josh Hawley told the DCNF that lawmakers should be concerned about superintelligence, but stopped short of endorsing a ban of the tech..

Kennedy introduced a bill Wednesday to require AI models to have a “kill switch,” which would allow companies to completely shut down any AI model if it acts amorally or immorally outside of human control. Republican Kentucky Sen. Rand Paul blocked its passage after Kennedy attempted to pass the bill without holding a formal vote, which just one senator can prevent.

Paul tried to amend Kennedy’s bill to establish a group that could make recommendations on AI guardrails to the Senate Commerce Committee. Kennedy rejected his suggestion.

President Donald Trump called warnings against AI a “sick conspiracy” and opposed any guardrails, warning that China would beat the U.S. in AI development if companies slowed down.