HR 1694, the AI Accountability Act, directs the Commerce Department's Assistant Secretary for Communications and Information to study accountability measures for AI systems used in communications networks (like social media and telecom) and spectrum sharing. The study will examine how these measures can help close the digital divide, reduce cybersecurity risks, and clarify terms like "trustworthy" AI, while consulting with industry, academia, and consumers. It requires two reports within 18 months: one detailing the study findings and stakeholder feedback, and another recommending how to make AI system information accessible to the public. The bill does not impose new regulations but focuses on gathering data to inform future policy.
This bill creates a federal private right of action for individuals whose "covered data" (including personal information, biometrics, location data, and inferred profiles) is used without their clear, upfront permission. It requires explicit consent for data collection, processing, or sharing with third parties, mandating specific disclosures separate from general terms. Individuals can sue for damages (minimum $1,000 per violation), punitive damages, or injunctions, and the law invalidates pre-dispute arbitration agreements for these claims. The bill does not override stricter state privacy laws but establishes a baseline federal standard for data misuse by AI systems and data processors.
The SUCCESS for BEAD Act (HR 6920) amends the BEAD program to allow states and other eligible entities to use remaining broadband grant funds for specific purposes. It authorizes the use of these funds for AI-supportive telecommunications infrastructure, workforce development, public safety, and national security initiatives through a competitive subgrant program. The bill establishes specific requirements including a public challenge process for fiber infrastructure projects and priorities for underserved areas and strategic infrastructure. It also includes detailed definitions and requirements for Next Generation 9-1-1 systems in emergency communications centers. The bill provides more specific guidance on how remaining funds can be used without changing the overall BEAD program structure.
S 305, the Small Business Technological Act of 2025, expands eligibility for Small Business Administration (SBA) Section 7(a) loans to cover business software, cloud computing services, and AI-powered tools that support core operations like payroll, HR, sales, billing, accounting, and inventory management. The bill directly affects small businesses seeking SBA loans by allowing these funds to finance technology tools they previously could not use for such purposes. It explicitly excludes research and development from eligible uses and clarifies that existing working capital definitions remain unchanged. This policy change modifies the SBA’s existing loan program without creating new funding or altering prior loan approvals for qualifying technology.
The READ AI Models Act directs the National Institute of Standards and Technology (NIST) to develop a free, flexible template for documenting artificial intelligence (AI) models. The template would require basic details like developer information, training data cutoff dates, supported languages, and terms of service for both public and private sector AI models. NIST must gather public input through a 60-day comment period before finalizing the template and technical guidelines. If the pilot program proves effective, NIST would publish the template publicly and report to Congress within one year.
This bill directs the Election Assistance Commission (EAC) to create voluntary guidelines for state and local election offices on managing artificial intelligence (AI) in elections. The guidelines must address AI's risks and benefits, cybersecurity threats, how AI affects sharing accurate election information, and how it might spread disinformation. The EAC must also study AI's use in the 2024 federal elections by July 2026 and update the guidelines based on that study. These guidelines are non-binding, meaning election offices choose whether to adopt them, and they aim to help election administrators navigate AI tools responsibly.
This bill requires the Department of Energy (DOE) and the National Oceanic and Atmospheric Administration (NOAA) to collaborate on research using advanced computing techniques - like AI, high-performance computing, and data analytics - to improve weather and climate models. It establishes a competitive funding process for federal agencies, National Laboratories, and universities to develop new models and optimize computing infrastructure, with up to three "centers of excellence" at National Labs prioritizing existing NOAA partnerships and AI computing expertise. The bill mandates reports to Congress within two years detailing progress, collaboration effectiveness, and future opportunities, and expires after five years. It directly affects NOAA, DOE, National Laboratories, and academic institutions participating in the funded research.
This bill creates a federal program requiring developers of powerful AI systems (defined as those using over 10²⁶ computing operations) to participate in mandatory testing. The Department of Energy will evaluate these systems for risks like loss-of-control scenarios, weaponization threats, or harm to civil liberties, collecting data on safety and behavior. Developers must share code, training data, and technical details with the program and face $1 million daily fines for non-compliance. The program runs for seven years, producing annual reports and a final oversight plan for Congress to inform future regulations on advanced AI.
This bill establishes the Defense Technology Hubs Program, requiring the Secretary of Defense to create 10 regional hubs by 2028. These hubs - formed by partnerships of universities, defense contractors, small businesses, and state/local governments - will focus on accelerating development of defense technologies like AI, quantum computing, and hypersonics. The program authorizes $375 million (2026-2030) for grants, with hubs using funds for research, prototyping, and workforce training while complying with security rules like export controls (ITAR/EAR) and cybersecurity standards. It prioritizes geographic diversity and aims to strengthen collaboration between defense agencies, industry, and academic institutions to boost national security innovation.
This bill amends the Federal Food, Drug, and Cosmetic Act to allow artificial intelligence (AI) and machine learning technologies to legally prescribe drugs under specific conditions. It requires AI systems to be both authorized by state law and approved by the FDA under sections 510(k), 513, 515, or 564. The law directly affects healthcare AI developers, providers, and patients by establishing a regulatory framework for AI-driven drug prescriptions. Key provisions define AI as a "practitioner" for prescription purposes only when meeting these two federal and state requirements. This creates a clear pathway for AI to legally prescribe drugs within existing FDA and state oversight.
