Who Pays for the Truth? The UK's Copyright Battle with Big Tech

# Who Pays for the Truth? The UK's Copyright Battle with Big Tech > The UK government has to make a choice: protect the creative industries or water down copyright law to attract US tech. We sit down with the FT's Global Policy Director to discover whats at stake. [Read on Substack](https://lawwhatsnext.substack.com/p/who-pays-for-the-truth-the-uks-copyright) · 2026-03-13 · Law What's Next --- 🎙️This week we publish our conversation with Matt Rogerson, Global Policy Director at the Financial Times, and one of the more prominent voices in the UK press and publishing industry on the question of AI companies using copyrighted content without permission or payment. The timing could hardly be more significant. We got together the day the House of Lords Communications and Digital Committee published what may prove to be the most consequential UK report on AI and creative industries to date; and, just one week after the FT helped launch SPUR, a coalition of the UK’s biggest news organisations fighting for fair AI licensing standards. But to understand why this conversation matters, let us try to set the scene on what provoked it… The State of Play: UK Copyright Law and AI What the law currently says Under Section 29A of the Copyright, Designs and Patents Act 1988, the UK allows text and data mining (TDM) only for non-commercial research purposes. If you want to mine copyrighted content for commercial purposes - such as training an AI model - you need a licence from the rights holder. What the government proposed In December 2024, the UK government launched a public consultation on “Copyright and AI”. It set out four possible options, ranging from leaving the law unchanged to allowing AI companies to use copyrighted material with no restrictions at all. Crucially, the government indicated a preferred option: introduce a broad commercial TDM exception with an opt-out mechanism for rights holders. In other words, AI companies would be free to scrape and use your content unless you actively told them not to. The creative industries pushed back hard. Over 400 of the UK’s leading media and arts professionals wrote to the Prime Minister. Musicians, authors, photographers and publishers warned that the proposal would gut the economic model that funds their work. Elton John called the government “absolute losers” over the prospect. The consultation received a wave of responses and crucially, none of the government’s proposed options achieved broad support. The “reset” moment In January 2026, Secretaries of State Liz Kendall (Science, Innovation & Technology) and Lisa Nandy (Culture, Media & Sport) appeared before the Lords Communications and Digital Committee. They conceded the government had been wrong to express a preference and described the moment as a “reset.” Nandy stated that the government would not rush into a decision, acknowledging: “If we rush into this and get it wrong, we could make a mess.” Under the Data (Use and Access) Act 2025, the government is required to publish an economic impact assessment and a report on its approach to copyright and AI by 18 March 2026 - that’s next week 👀 The Lords Committee Report: a line in the sand On 6 March 2026, the House of Lords Communications and Digital Committee published its full report: AI, Copyright and the Creative Industries. It is 85 pages long, draws on testimony from Google, Meta, Microsoft, OpenAI, and dozens of creative industry bodies, and its conclusions could not be clearer. The committee found that the UK’s copyright framework is not outdated or in need of reform. The problems stem from widespread unlicensed use of protected works and a lack of transparency from AI developers about their training data. It recommended that: The government rule out a commercial TDM exception entirely Make transparency about AI training data a statutory obligation Introduce protections against unauthorised digital replicas and “in the style of” AI outputs Support the development of sovereign AI models The committee laid out two possible futures for the UK: Becoming a world-leading home for responsible, licensing-based AI development; or Drifting toward acceptance of large-scale use of unlicensed creative content by opaque US-based AI models. In 2023, the UK’s creative industries contributed £124 billion to the economy and employed 2.4 million people. The AI sector as a whole contributed £11.8 billion and employed around 86,000. The committee made clear that trading the former for speculative gains from the latter could be a poor bet. Share The Lobbying Battle The tech side AI companies and their trade bodies have taken what Matt describes as “an extreme position” which, he notes, is what you do in any negotiation. The argument runs along familiar lines: AI is transformative, the UK risks falling behind competitors, and copyright enforcement at inference time would reduce AI usefulness and make UK businesses less competitive than those in countries with more permissive regimes. A lot of money is being spent to advance this argument. The pamphlet titled “Text & Data Mining and its value to the UK economy” that triggered Matt’s most recent LinkedIn outburst and this conversation was published by Public First, a UK policy consultancy. It called for a broad commercial TDM exception in UK copyright law and extended the argument to cover AI inference (more on this below) as well. Its core claims are that enforcing copyright during inference is an extreme position, that doing so would reduce the usefulness of AI services, and that UK businesses would get worse results than competitors in more permissive jurisdictions. The creative side The creative industries have organised with increasing force. In addition to the open letters, Parliamentary hearings, and media coverage, the sector has moved from protest to action. One such significant development is SPUR — the Standards for Publisher Usage Rights coalition — launched on 26 February 2026 by the Financial Times, BBC, The Guardian, Sky News and The Telegraph. SPUR is not a collective licensing body and will not set fixed prices. Instead, it aims to establish shared technical standards and licensing frameworks that allow AI developers to access quality journalism through rights-cleared channels, while ensuring publishers retain control and receive fair value. Its founding members are evaluating pricing models from “pay-per-crawl” to “pay-per-inference” and the coalition is open to global publishers. The Conversation with Matt Rogerson Against that backdrop, our conversation with Matt cuts to the commercial and philosophical heart of the debate. Here are the key arguments Matt makes. The anthropomorphisation trick Matt’s sharpest criticism is aimed at the rhetorical strategy used by tech companies to normalise unlicensed content use. They anthropomorphise AI models — drawing parallels between how machines process information and how humans learn — as a way to argue that the same fair use principles should apply. As Matt puts it, this tactic “moderately enrages” him. Machines are not humans. They copy at scale, at speed, for profit. The analogy is convenient, but it doesn’t hold. Inference is where the real value lies Most of the copyright debate to date has focused on training, i.e. the process of scraping content to build models. But the Public First pamphlet pushes into inference territory: that being what happens when a deployed model actively retrieves or references content to answer a user’s query in real time. This matters enormously. As AI systems evolve toward retrieval-augmented generation and agentic workflows, they increasingly access current, real-time content at inference time. This is not a one-off historical act of learning. It is an ongoing, commercial use of copyrighted material every time a user asks a question. Training builds the engine. Inference drives the car using someone else’s fuel. The infrastructure already exists Matt highlights one of the more frustrating arguments in the broader debate, that is the implication that it is simply too hard to build mechanisms for AI companies to pay for content they use. Matt dismantles this comprehensively, citing the fact that: Microsoft launched its Publisher Content Marketplace in February 2026 — a two-sided marketplace where publishers set their own licensing terms and AI developers pay based on usage. Its launch partners include The Associated Press, Business Insider, Condé Nast, Hearst, USA Today and Vox Media. The FT already sells paid API access to its journalism. Prominent technologists such as Florent Daudens have set out a clear vision for news publishers in a post-browser/agentic future.. The claim that licensing is impractical is not just wrong — it is being actively disproved by the market. Our conversation with Matt is also available on Spotify, Apple Podcasts, or wherever you enjoy your podcasts. The sovereign AI opportunity Matt makes a compelling case that the UK does not need to give away its creative heritage to participate in the AI revolution. He points to the Allen Institute in the US — a model co-funded by the US government and Nvidia — which is completely open and transparent about its underlying training data. Institutions can host it on their own infrastructure, understand exactly how outputs are generated, and avoid dependency on any single hyperscaler. The House of Lords Communications and Digital Committee report reinforces this, recommending the UK prioritise sovereign AI models that deliver enhanced transparency and respect for copyright. The slop spiral Our exchange towards the end of the conversation touches on what may be the most consequential long-term argument. If there is no economic incentive to produce high-quality journalism — because AI companies can take it for free — then the supply of reliable information will inevitably degrade. AI models which depend on fresh, accurate content to function, are then trained on and will retrieve from an increasingly polluted information environment. The outputs will get worse, and trust will erode. We may then enter a “dark world” of AI-generated slop: content that is neither impartial nor trustworthy, dependent wholly on the alignment of a particular model and the commercial interests of those administering it. Matt highlights the potential for secure news and information supply chains becoming a national security issue if this dynamic starts to accelerate! Copyright://WhatsNext The government faces a deadline of 18 March 2026 to publish its economic impact assessment and report on copyright and AI. Meanwhile, the market is not waiting for the government. Microsoft is building its marketplace. SPUR is building its coalition. Publishers are striking deals. And the Lords Committee has handed the government a comprehensive, evidence-based case for a licensing-first approach. The question is whether the government will take it — or whether the lobbying power of the tech incumbents will persuade ministers that giving away the UK’s creative output is a price worth paying for a seat at the AI table. As the Chair of House of Lords Communications and Digital Committee, Baroness Keeley said: “AI may contribute to our future economic growth, but the UK’s creative industries create jobs and economic value now... Watering down the protections in our existing copyright regime to lure the biggest US tech companies is a race to the bottom that does not serve UK interests.“ We agree. Enjoy your weekends (when they arrive)! Tom & Alex Thanks for reading Law://WhatsNext! Subscribe for free to receive new posts and support my work.