The Trump administration on Friday dropped a skimpy artificial intelligence (AI) policy framework on Congress that beseeches lawmakers to establish a single national standard that would strip states of their power to regulate the booming technology.
The proposal aims to cement American dominance in the global AI race as well as address rising public concerns over child safety, energy costs, and digital rights.
The four-page outline marks the administration’s most aggressive move yet to halt what it calls a “patchwork” of state-level regulations. California, Colorado, Texas, and Utah have enacted their own AI laws, focusing on consumer transparency and data privacy. However, White House AI czar David Sacks argued that 50 discordant regulatory regimes “threaten to stifle innovation and jeopardize America’s lead.”
The administration’s America First approach to AI seeks to provide the tech sector — led by giants like NVIDIA Corp., Microsoft Corp., and Alphabet Inc. — with the regulatory certainty needed to outpace global rivals like China. To enforce this vision, President Trump previously threatened to withhold federal broadband funding from states whose local AI laws are judged to hinder national technological progress.
Republican leaders, including House Speaker Mike Johnson, praised the roadmap for prioritizing “innovator certainty.” A key pillar of the plan involves streamlining permits for data centers, allowing the energy-intensive facilities to generate their own power on-site to prevent surging electricity costs for American families.
Although the policy is heavily pro-innovation, it seeks a “bipartisan consensus” by focusing on online safety. The framework urges Congress to give families direct control over accounts and devices to protect children’s privacy; implement features to prevent AI-generated self-harming content and sexual exploitation; and include strict provisions to prevent AI-driven censorship and protect free speech.
The administration also waded into the contentious legal battle between AI developers and creators. The framework explicitly states the administration’s belief that training AI models on copyrighted material “does not violate copyright laws,” aligning with “fair use” arguments favored by tech firms. However, acknowledging the $1.5 billion settlement recently reached between Anthropic and various authors, the White House suggested that courts should continue to resolve specific disputes.
Despite the administration’s push to codify this framework into law this year, the path through a fractured Congress remains uncertain. Republican majorities are currently focused on the SAVE America Act and other pre-midterm priorities.
Further, the policy notably downplays certain national security risks. While China hawks in Washington have raised alarms over hardware exports, the administration recently greenlit shipments of NVIDIA’s second-most advanced chips to China.
Sacks defended the move, arguing that supplying American chips discourages rivals like Huawei from developing their own advanced architectures.
The White House stated it intends to work with lawmakers in the coming months to transform these principles into a formal bill for the President’s signature.
