US AI Policy News Today: Key Updates & Government Actions

Source: Dev.to

US AI policy news today features a flurry of government action across multiple fronts. Policymakers are scrambling to build America’s AI advantage while setting guardrails – almost like trying to catch a rocket after it has launched. In recent months, the US government has rolled out executive orders, new initiatives, and legislation around artificial intelligence, aiming to stay competitive without ignoring safety.

The picture is complex: some actions aim to sprint ahead on innovation, while others emphasize caution and risk management.

US AI Policy Report Card: Leadership vs Caution

Federal AI policy remains very much a work in progress. The US has no single AI law; instead it relies on a patchwork of executive actions and guidelines.

• January 2025 – The Trump administration issued an executive order titled “Removing Barriers to American Leadership in AI” (source: Federal Register). The order rescinded many of President Biden’s previous AI directives and instructed agencies to eliminate rules seen as hindering innovation.
• July 2025 – The White House published America’s AI Action Plan, a comprehensive strategy listing over 90 federal initiatives to boost U.S. AI development and leadership.
• October 2023 – President Biden signed an order on Safe, Secure, and Trustworthy AI (EO 14110) to promote ethical development.
• January 2025 – Biden issued an order on Advancing U.S. Leadership in AI Infrastructure, calling for domestic AI data centers powered by clean energy and modernized computing infrastructure.

These swings reflect different philosophies. Experts warn that deregulating AI alone won’t automatically deliver great results. Arati Prabhakar and Asad Ramzanali note that we need government‑led R&D to solve big problems (like rare diseases or education), not just unregulated chatbots. In their words, “we need clear‑eyed action to harness AI’s benefits,” not merely letting tech companies run wild.

Major Federal Initiatives and Bills

• Genesis Mission (Nov 2025) – A Trump‑era executive order likened to the Manhattan Project. It tasks the Department of Energy with creating an integrated AI research platform that leverages the nation’s federal science datasets to accelerate breakthroughs in energy, healthcare, national security, and more.
• AI Talent Act (Dec 2025) – A bipartisan bill (Rep. Sara Jacobs & Sen. Andy Kim) to help federal agencies recruit and retain top AI experts by creating specialized talent teams and streamlined hiring tools.
• AI Training for National Security Act (FY 2026 NDAA) – Requires the Pentagon to add AI and cyber‑threat content to basic training for troops and civilian staff.
• Executive Orders – Biden’s 2023‑2025 orders focused on safety and infrastructure; Trump’s 2025 orders pivot to boosting innovation and R&D.
• Congressional Legislation – The National AI Initiative Act (2020) funds R&D; new proposals like the AI Talent Act and NDAA provisions strengthen the AI workforce.
• R&D Funding – Significant new programs at DOE, NSF, and under the CHIPS Act channel billions into AI compute and research.
• Agency Guidance – FTC, Commerce, and other agencies have released guidelines on AI fairness, privacy, and safety; federal hiring and ethics policies are being updated.

Overall, the federal strategy mixes aggressive investment in innovation (e.g., the AI Action Plan) with selective oversight signals (e.g., the Safe AI EO). Analysts note that U.S. companies largely operate under existing laws, adapting voluntarily rather than facing brand‑new AI‑specific rules. Yet with dozens of new initiatives, the government is clearly upping its AI game.

State vs. Federal: A Patchwork Landscape

With no national AI law, states have rushed in. As of late 2025, over 45 states considered AI legislation and about 31 enacted some regulations.

• Colorado – Passed the nation’s first AI bias law for “high‑risk” systems (e.g., hiring and lending).
• California – Has dozens of pending AI bills covering content labeling, deepfakes, data privacy, and more.

The Trump administration responded in December 2025 by announcing an executive order that would block state AI regulations, arguing “there must be only one rulebook if we are going to continue to lead in AI.” Critics warn this deregulatory push could let tech companies evade accountability, while supporters say it avoids a confusing array of 50 different laws. South Dakota’s Attorney General emphasized that states should retain the right to impose “reasonable” AI regulations.

• Federal stance – Voluntary guidelines and agency enforcement (FTC, DoC, etc.), no sweeping AI law yet.
• State activity – A mosaic of laws on bias, privacy, content labeling, etc. (Colorado’s AI Act, California proposals, etc.).
• Tension – Trump’s proposed order would override state AI rules, drawing pushback from states that insist on retaining regulatory authority.

In everyday terms, it’s as if we wrote 50 separate rulebooks for AI (one per state) and are now debating whether a single unified manual would be simpler.

Industry and Emerging Voices

These policy shifts are unfolding alongside rapid industry changes. For example, AMD has been landing major AI contracts and building next‑generation AI supercomputers, pushing its data‑center revenue sharply upward. While AMD’s rise is primarily a business story, it ties into national strategy: U.S. policy favors a strong domestic AI hardware base. In the software world, companies like OpenAI, Google, and Microsoft continuously update their AI offerings (e.g., Copilot tools) and of…