On March 20, the Trump administration submitted a National Policy Framework for Artificial Intelligence to Congress. It is not an executive order and does not take effect on its own, but it marks the first time the federal government has formally stated how AI should be regulated, by whom, and on what terms.
The conclusion? Light regulation, federal uniformity, and a green light for industry.
Six Goals, With the Third Being Key
The framework outlines six policy directions:
- Minor protection: Parents should have control, platforms must implement age verification, and children's access to harmful content should be restricted.
- Community infrastructure: Ease approval for data centers, allow self-built power plants, but electricity costs must not be passed on to ordinary residents.
- Intellectual property and creators: Protect creators while preserving fair use space for AI training.
- Free speech: AI must not suppress legitimate political expression, and the government must not use AI to control public opinion.
- Innovation and U.S. leadership: Remove outdated regulations that hinder innovation and accelerate AI deployment across industries.
- Workforce development: Skills training for the AI era so workers can also share in growth dividends.
The six points sound balanced, but the real impact lies in the third point and the overall tone.
Core Move: Federal Uniformity, States Step Aside
The most interesting part of this framework is not what it says, but what it seeks to dismantle — the AI regulations that states have already passed or are advancing.
Over the past two years, U.S. states have been proliferating AI regulations: more than 600 AI-related bills are moving through various legislatures. New York's RAISE Act took effect on March 19, imposing transparency and safety requirements on large model companies; California, Colorado, and Texas are also developing their own frameworks.
The federal government's position is now clear: We will handle this uniformly; you should not each go your own way.
The framework states: "Fragmentation of state regulations undermines America's innovation competitiveness." AI model development is inherently interstate and should fall under federal jurisdiction; developers should not bear legal liability for third-party misuse.
For companies like OpenAI and Anthropic, this is great news. They have lobbied for federal uniformity for years, with a practical argument: 50 states with 50 sets of rules make compliance costs more expensive than building the models themselves. Now the White House is backing them.
Copyright Issue Directly Characterized
In the intellectual property section, the framework includes a controversial statement: it confirms that AI companies training models on copyrighted content falls under "fair use" and does not violate copyright law.
At the same time, the framework says it will protect creators' unique identities and digital likenesses from misuse and recommends that Congress establish protections against AI-generated fake portraits and voices.
It addresses both sides, but for copyright holders currently in court, the White House's characterization is glaring. The core dispute in lawsuits by The New York Times and music publishers against AI companies is whether training on copyrighted content is legal. While the White House's policy stance is not a ruling, it will influence the Justice Department's litigation position and affect judges and lawmakers' judgment.
The reaction from the creator community is predictable.
Is This AI Regulation or an AI Exemption Order?
To be honest, the "regulatory" element of this framework is very thin; it looks more like a protective umbrella tailored for the AI industry.
It opposes creating new AI regulatory agencies, advocates relying on existing institutions; opposes fragmentation of state-level regulations; sides with AI companies on copyright; and pushes for faster approval of data centers. Critics ask: Who will address AI safety risks? Who will ensure AI does not discriminate against vulnerable groups? Who will manage AI-generated misinformation?
The framework does not directly answer these questions.
Supporters will say: This is a race with China; excessive regulation would cede first-mover advantage. Loosening controls and accelerating deployment is the correct strategic choice.
Both logics have merit, but this framework is ultimately just a recommendation — whether it can be implemented depends on whether Congress is willing to follow up with legislation. The U.S. federal government still lacks a comprehensive AI law. Trump's framework is essentially urging Congress: Act quickly, before state laws run ahead.
In the coming months, the battle between federal and state-level AI legislation may be more intense than any model release.
Sources: CocoLoop, President Donald J. Trump Unveils National AI Legislative Framework (White House); White House Releases National Policy Framework for Artificial Intelligence (WilmerHale); The White House National Policy Framework for Artificial Intelligence: what it means and what comes next (Consumer Finance Monitor)