This bill establishes a moratorium on constructing or upgrading new artificial intelligence data centers until specific federal laws are enacted to ensure AI safety, protect workers from job displacement, and prevent environmental harm. It defines AI data centers as facilities with significant power capacity or advanced cooling systems used for large-scale AI model development. The legislation requires the Secretary of Energy to submit quarterly public reports on data center operations, including details on energy use, emissions, water consumption, and labor practices. Additionally, it prohibits the export of computing infrastructure hardware to countries that lack comparable AI safety regulations or to entities using such hardware for large-scale AI training and deployment.
The GUARDRAILS Act repeals a December 2025 executive order on artificial intelligence policy and prohibits federal funding for its implementation. This legislation directly affects federal agencies by removing their authority to enforce the previous executive order's framework. The bill does not create new AI regulations but instead eliminates the existing executive order that established a national policy framework for artificial intelligence.
The GUARDRAILS Act (HR 8031) repeals the December 11, 2025 Executive Order on Artificial Intelligence, which previously established a national policy framework for AI development. By removing this executive order, the bill prevents the federal government from using funds to implement, enforce, or administer the policies outlined in that directive. This change directly affects federal agencies and any organizations that were relying on the executive order's framework for AI regulation. The legislation does not create new AI rules but instead eliminates the existing executive mandate that had been in place.
This bill, titled the AI Guardrails Act of 2026, establishes specific restrictions on how the Department of Defense can use artificial intelligence. It prohibits the military from using AI to launch or detonate nuclear weapons, to monitor or target individuals within the United States without a legal basis, and to employ lethal force through autonomous weapon systems without human oversight. The legislation also creates a waiver process that allows the Secretary of Defense to temporarily permit certain autonomous weapon uses under strict conditions, requiring written certification to Congress and detailed notifications about the system's safeguards and testing results.
HR 1122, the China Technology Transfer Control Act of 2025, requires the U.S. President to control exports of specific technologies and intellectual property to China after 180 days of enactment. It targets "covered national interest technology" defined as items aiding China's military, supporting its Made in China 2025 industrial policy (like semiconductors or AI), or used in human rights violations. The bill mandates sanctions against foreign entities that knowingly transfer such technology to China, with exceptions for imported goods. It also directs the U.S. Trade Representative to create an annual list of China-produced goods tied to these policies or human rights abuses.
HR 4894, the Deceptive Practices and Voter Intimidation Prevention Act of 2025, prohibits the intentional spread of false information about voting procedures, eligibility, or penalties within 60 days of an election. The bill specifically targets deceptive communications through social media, text messages, and AI-generated content designed to prevent voters from casting ballots, including false claims about voting locations, registration status, or legal consequences. It creates a private right of action for individuals harmed by these deceptive practices and authorizes the Attorney General to issue corrective information to counter false claims. The law directly affects voters, election officials, social media platforms, and anyone spreading false voting information, with penalties including fines and imprisonment for violations.
HR 6624, the Biological Intellectual Property Protection Act of 2025, requires export licenses for digital files representing human- or AI-designed synthetic DNA or RNA sequences when sent to foreign entities linked to countries designated as national security concerns (like China). It directly affects biotechnology companies and researchers exporting such digital sequence data, mandating licenses for these specific exports within one year of the law's enactment. The key mechanism is a new license requirement under export controls, targeting digital files that could reveal sensitive biotechnology research. This policy change aims to prevent foreign adversaries from accessing U.S. biotech intellectual property through synthetic biology data exports.
This bill requires that public comments on federal land management rules (administered by the Bureau of Land Management) must come only from U.S. citizens, replacing the previous standard. It mandates that agencies implement CAPTCHA systems to prevent artificial intelligence from submitting comments, explicitly stating this is to deter AI-driven public involvement. The changes apply to rulemaking under the Federal Land Policy and Management Act, affecting how the public can participate in decisions about national public lands. The bill does not alter land management policies but modifies the process for gathering public input.
This bill prohibits U.S. federal agencies from purchasing large language models (LLMs) that do not meet specific standards for truthfulness, neutrality, and accuracy. It requires all federal LLM procurement after enactment to prioritize factual responses, historical/scientific accuracy, and nonpartisan outputs - explicitly banning models that encode partisan views or favor concepts like diversity, equity, and inclusion unless prompted by users. The rule applies to all federal agencies (excluding the Government Accountability Office) and directly affects LLM developers seeking government contracts. Key provisions mandate that LLMs must acknowledge uncertainty, avoid ideological manipulation, and remain neutral in responses. The law creates a concrete procurement standard for federal AI tools, not a broader regulation.
HR 6361, the Ban AI Denials in Medicare Act, prohibits the U.S. Department of Health and Human Services from implementing or testing Medicare models that use artificial intelligence for prior authorization of covered services under traditional Medicare. Specifically, it blocks the existing WISeR model (described in a 2025 federal notice) and prevents future models from using AI to deny coverage for Medicare Part A or Part B services. This bill directly affects Medicare beneficiaries and healthcare providers who would otherwise face AI-driven coverage decisions. The key mechanism is an amendment to the Social Security Act, explicitly banning the use of AI in prior authorization systems within Medicare's innovation models.
