White House releases AI legislation framework
The White House released a new framework for domestic AI legislation Friday morning, focused on protecting children and strengthening the industry, while calling for strict limits on developers’ legal liability and state laws that it said would slow the technology’s development.
The legislative proposal underscores the need for Congress to establish a unifying federal approach to artificial intelligence rather than letting states set individual rules that it fears could hinder innovation, a position the White House has repeatedly signaled in recent months. Politicians and activists across the political spectrum have instead argued for states’ ability to regulate AI in the absence of significant federal action, while Congress debates how to regulate this rapidly evolving technology.
“The federal government is in a unique position to set a coherent national policy that allows us to win the AI race and deliver it to the American people,” the White House said in a statement accompanying the release of the framework.
“The administration looks forward to working with Congress in the coming months to transform this framework into law the president can sign.”
The framework is divided into seven main areas, ranging from “Protecting Children and Empowering Parents” and “Respecting Intellectual Property Rights and Supporting Creators” to “Educating Americans and Developing an AI-Ready Workforce.”
Several of the framework’s provisions, including a focus on protecting children and supporting building a U.S. AI infrastructure, were previewed in President Donald Trump’s December executive order. That order directed David Sacks, the White House AI chief, and Michael Kratsios, director of the Office of Science and Technology Policy, to create Friday’s draft framework.
The framework supports limiting developer liability for harm caused by AI systems, particularly against “unlimited liability” that “could give rise to excessive litigation” for matters related to child safety. The framework also puts forward limits on the ability of states to “penalize AI developers for illegal third-party behavior involving their models.”
These proposed liability restrictions align with messages from Sacks, a venture capitalist, and many high-profile Silicon Valley investors, who say significant liability provisions would harm U.S. AI innovation and scare away future investment.
The need to regulate America’s burgeoning AI industry has quickly become a uniting factor for MAGA conservatives and progressive activists. In recent months, slowing the spread and construction of data centers has become a key bipartisan issue in many state capitals.
Although there is currently no sweeping federal legislation governing AI, California’s SB 53 and New York’s RAISE Act have set the current standard for AI legislation. Both laws require major AI companies, such as OpenAI, Anthropic, and Google, to implement additional whistleblower protections for employees, report any significant security-related events to state offices, and disclose how they test their models for key risks.
The Trump administration’s calls to restrict states’ ability to legislate on AI have angered many Republicans recently. In a letter to Trump in early March, more than 50 Republicans said that “recent attempts to end domestic AI legislation suggest not only a desire for coordination, but also an effort to prevent the adoption of measures holding the tech industry accountable.”
The letter was released in response to the Trump administration’s pressure campaign against a proposed bill in Utah that would require AI companies to be more transparent about how they work to protect children using their systems and how they plan to limit the catastrophic risks of their models, such as helping terrorists create biological weapons or crippling cyberattacks.
The framework proposed Friday says states should retain the authority to pursue matters that would normally fall within their jurisdiction, such as fraud prevention and consumer protection.
The policy document also prioritizes protections against AI-related censorship, advocating that Congress prevent the federal government “from compelling technology providers, including AI providers, to ban, coerce, or edit content based on partisan or ideological agendas.”
