Lord Clement-Jones (LD)

My Lords, I declare my AI interests as set out in the register. I will speak in support of Amendments 204, 205 and 206, which have been spoken to so inspiringly by the noble Baroness, Lady Kidron, and so well by the noble Lords, Lord Freyberg, Lord Lucas and Lord Hampton, the noble Earl, Lord Clancarty, and the noble Viscount, Lord Colville. Each demonstrated different facets of the issue. I co-chair the All-Party Group on AI and chaired the AI Select Committee a few years ago. I wrote a book earlier this year on AI regulation, which had a namecheck from the noble Baroness, Lady Jones, at Question Time, which I was very grateful for. Before that, I had a career as an IP lawyer, defending copyright and creativity, and in this House, I have been my party’s creative industries spokesperson. The question of IP and the training of generative AI models is a key issue for me. This is the case not just in the UK but around the world. Getty and the New York Times are suing in the United States, as are many writers, artists and musicians. It was at the root of the Hollywood actors’ and writers’ strikes last year. It is one thing to use the tech—many of us are AI enthusiasts—but it is another to be at the mercy of it. Close to home, the FT has pointed out, using the index published by the creator of an unlicensed dataset called Books3, published online, that it is possible to identify that over 85 books written by 33 Members of the House of Lords have been pirated to train AI models from household names, such as Meta, Microsoft and Bloomberg. Although it is absolutely clear that we know that the use of copyrighted works to train AI models is contrary to UK copyright law, the laws around the transparency of these activities have not caught up. As we have heard, as well as using pirated e-books in their training data, AI developers scrape the internet for valuable professional journalism and other media, in breach of both the terms of service of websites and copyright law, to train commercial AI models. At present, developers can do this without declaring their identity, or they may use IP scraped to appear in a search index for the completely different commercial purpose of training AI models. How can rights owners opt out of something that they do not know about? AI developers will often scrape websites or access other pirated material before they launch an LLM in public. This means that there is no way for IP owners to opt out of their material being taken before its inclusion in these models. Once used to train these models, the commercial value, as we have heard, has already been extracted from IP scraped without permission, with no way to delete data from these models. The next wave of AI models responds to user queries by browsing the web to extract valuable news and information from professional news websites. This is known as retrieval-augmented generation—RAG. Without payment for extracting this commercial value, AI agents built by companies such as Perplexity, Google and Meta will, in effect, free-ride on the professional hard work of journalists, authors and creators. At present, such crawlers are hard to block. There is no market failure; there are well-established licensing solutions. There is no uncertainty around the existing law; the UK is absolutely clear that commercial organisations, including gen AI developers, must license the data that they use to train their large language models. Here, as the Government’s intentions become clearer, the political, business and creative temperature is rising. Just this week, we have seen the creation of a new campaign, the Creative Rights in AI Coalition—CRAIC —across the creative and news industries and, recently, Ed Newton-Rex reached more than 30,000 signatories from among creators and creative organisations.