AI policy on the copyright front is taking shape across the globe, with emerging signs of a transatlantic split. In the United Kingdom, the government recently retreated from its preferred artificial intelligence training opt-out model after industry pushback and now lacks a settled position. In the U.S., the Trump Administration newly-issued framework on AI takes a more assertive approach – stating that training AI on copyright-protected material does not violate copyright law, while leaving the fair use question for the courts.
This points to an emerging divergence in how major economies are approaching AI training, intellectual property, and creator protections.
The United Kingdom: Retreat Without Resolution
This month, the UK government withdrew support for its 2024 proposal to introduce a broad copyright exception allowing AI developers to train on copyright-protected material subject to an opt-out. The proposal drew strong criticism from artists, publishers, and industry groups, who argued it would weaken copyright protections and enable uncompensated use of creative works.
Following consultation, the government concluded in its March 18 report and written statement that the opt-out model is no longer its preferred approach and that it now has no preferred option for reform, underscoring the lack of consensus on how to balance competing objectives. It has framed the issue as a trade-off between ensuring creators can control – and be compensated for – their work and enabling AI developers to access training data.
The government is now considering additional measures, including greater creator control over online use, protections against unauthorized digital replicas, and potential labeling requirements for AI-generated content. It has not ruled out new copyright exceptions for AI training, leaving the overall direction unsettled.
> For now, the UK’s position amounts to a pause: its earlier model has been rejected, but no replacement has emerged.
The United States: Judicial Resolution & Limited Intervention
Meanwhile, the U.S. has articulated a clearer – though still evolving – position in the White House’s March 2026 legislative framework on artificial intelligence. The framework states that the Trump Administration “believes that training of AI models on copyrighted material does not violate copyright laws,” but also acknowledges that arguments to the contrary exist. The White House states that it “therefore supports allowing the courts to resolve this issue,” and says that Congress “should not take any actions that would impact the judiciary’s resolution of whether training on copyrighted material constitutes fair use.”
Rather than legislating the issue, the framework defers judicial determination of whether training constitutes infringement or fair use to the courts. It also outlines complementary measures to address creator concerns …
> Licensing and collective rights mechanisms: Congress should consider enabling frameworks that allow rights holders to collectively negotiate compensation from AI providers without violating antitrust law, while not determining when or whether licensing is legally required.
> Digital replica protections: The framework proposes federal protections against the unauthorized distribution or commercial use of AI-generated replicas of an individual’s voice, likeness, or other identifiable attributes, subject to First Amendment safeguards.
> Ongoing monitoring: Policymakers are encouraged to monitor how courts develop copyright doctrine in the AI context and to consider additional legislative action only if gaps emerge.
Taken together, this reflects a strategy that prioritizes innovation and legal flexibility, while relying on existing judicial processes to define the boundaries of copyright law as applied to AI.
Implications for Creators & AI Developers
For creators, the UK’s decision to step back from the opt-out proposal avoids an immediate shift toward a more permissive regime for AI training. However, the absence of a clear alternative leaves ongoing uncertainty about how their rights will be protected in the long term.
In the U.S., AI developers benefit from a policy environment that does not impose new statutory restrictions on training practices in the near term. At the same time, significant legal uncertainty remains, as courts will ultimately be tasked with determining whether and under what conditions such practices are lawful.
THE BOTTOM LINE: The UK and U.S. are confronting the same core issue – how to balance intellectual property protection with AI innovation – but are beginning to move in different directions. The UK has stepped back without settling on a replacement, while the U.S. is relying on courts and targeted measures rather than immediate legislation. This emerging divergence raises the prospect of regulatory fragmentation, potentially complicating compliance for companies operating across jurisdictions, even as policy frameworks continue to evolve.
Share
