OpenAI to UK: Adopt broad data mining exception or lose AI race

OpenAI has thrown its weight behind a sweeping text and data mining exception in UK law, telling Parliament that anything less will cede the country’s ambition to become Europe’s AI capital. The submission, part of the Government’s ongoing AI and copyright consultation, backs what’s known as Option Two—a broad TDM carve-out that would let AI firms train models on copyrighted material without needing individual licenses.

The company’s argument is stark. Laws are national but technology is borderless, and the UK is competing with jurisdictions that already provide clearer paths for AI development. OpenAI points to British researchers’ foundational role in AI and the country’s high adoption rates as strengths, but insists those advantages evaporate without a regulatory framework that treats data access as critical infrastructure. “Innovation and investments in AI infrastructure are only possible in jurisdictions where laws clearly support technological research and development,” the submission states.

OpenAI takes a direct swipe at the European Union’s approach, calling its opt-out system a mess of unclear technical standards that has created uncertainty for everyone involved. The contrast they draw is with the United States, where a more permissive environment has, in their view, sustained American dominance through multiple technology waves. The company argues a broad exception isn’t a zero-sum game—it can be designed to address specific rightsholder concerns while still fueling the kind of breakthroughs in education, science, and healthcare that policymakers claim to want.

The framing leaves no room for ambiguity. The UK either creates a predictable environment that attracts AI investment and talent, or it watches other jurisdictions shape the technology’s future. For a government that has made AI leadership a stated priority, the message from one of the field’s most prominent players amounts to a policy ultimatum.