Where UK AI Policy Is Headed: Insights from the House of Lords
AI policy debates don’t usually unfold in centuries-old chambers but last week, they did. I had the privilege of sitting in on a House of Lords session titled “AI Regulation, Ethics and Governance: Where Are We Now, and Where Are We Heading?”, co-chaired by Lord Clement-Jones CBE and Lord Kulveer Ranger. The room was packed with heavyweight voices from law, government, and tech, all focused on a single question: how should the UK shape its AI future? From patents for neural networks to data sovereignty and the ethics of open-source models, the discussion didn’t just skim the surface, it tackled the tensions and trade-offs head-on.
Dr. Rachel Free on Neural Network Patents and Data Sovereignty
Dr. Rachel Free, Partner and Patent Attorney at CMS Tax Law, reminded us that the UK Supreme Court will deliver its July judgment in Emotional Perception vs Comptroller General of Patents - a landmark case on whether patents designed for computer programmes can apply to artificial neural networks. While lower courts sided with the IPO, the High Court granted an appeal, and next month’s ruling is due to set a binding precedent.
Dr. Free also addressed data sovereignty, citing DeepSeek, a startup that built a foundation model entirely on analysis of other top-tier models, to demonstrate that it is both feasible and practical to keep training datasets, IP rights, and model control on-shore rather than solely relying on foreign providers.
Chris Mills on the IPO’s Approach to AI
Chris Mills, Director of Rights Policy at the UK Intellectual Property Office (IPO), stressed that although Parliament first considered AI back in 1987, the field has never been more pressing than at present.
He outlined three strategic objectives for the ongoing copyright and artificial intelligence consultation:
Strengthen creators’ rights in regards to data scraping and AI emulation.
Enable home-grown AI innovation by clarifying patent and copyright rules.
Hold AI firms accountable for responsible data use and downstream impacts.
In response to a question from the audience regarding a National Data Library (a curated, privacy-preserving repository of public-sector and licensed private data) to replace indiscriminate web scraping, Chris highlighted that such a scheme is included in the government’s AI action plan. He agreed that it would boost transparency in the sector and ensure that training resources are ethically sourced, while bringing clarity to the creative sector.
Chris Yiu on Open-Source Models
Chris Yiu, Director of Public Policy, Northern Europe at Meta Platforms, was direct: “Clarity is key”. He was complimentary of the government’s steps so far, and urged both houses of Parliament to continue to constantly engage with industry. Chris was particularly passionate about Meta’s open approach to its cutting-edge AI models, stating that its disclosure of model weights enables businesses to align the models they use with their unique company culture.
He also faced difficult questions from the crowd; in response to a question concerning perceived biases in the model, he stressed Meta’s leadership in developing open systems since the company’s founding.
Dr. John Fletcher on Openness, Innovation and Exponential Lead
Dr. John Fletcher, CEO of The Innovation Game, drove home a stark truth: “AI has joined humans as a peer in innovation”. From novel drug candidates to brand-new mathematical proofs, today’s models can originate ideas and then refine themselves through successive training cycles.
However, a frontier threat is emerging: self-improvement loop risks amplifying the lead of any closed, proprietary model into an unassailable market monopoly. Dr. Fletcher argued that open-model frameworks are the antidote: if leading models publish their architectures, training-data provenance and performance benchmarks, competitors have a chance to challenge them.
Achieving Global Standards for AI
The UK’s first Minister for AI, Viscount Camrose, asked a pointed question about the futility of nationally-siloed regulations and standards for AI. He advocated for support of global standards for AI, that align with developments in the sector and build international consensus.
Although the panel agreed with the Viscount’s advocacy for global standards, they also highlighted that the UK’s domestic policy in training and applications for AI has the potential to shape AI development, and certainly domestic AI use.
As we await the Supreme Court’s landmark ruling next month and the IPO’s copyright guidance, one thing is clear: the UK’s competitiveness hinges on balancing pro-innovation openness with robust, transparent governance. The choices we make today in data practices, patent frameworks and preference for open-source will define Britain’s role as a global AI leader for decades to come.
