Lord Vallance of Balham (Lab)

I congratulate the noble Baroness, Lady Kidron, on her excellent speech. I know that she feels very strongly about this topic and the creative industries, as do I, but I also recognise what she said about junior Ministers. I have heard the many noble Lords who have spoken, and I hope they will forgive me if I do not mention everyone by name. It is vital that we get this right. We need to give creators better, easier and practical control over their rights, allow appropriate access to training material by AI firms and, most importantly, ensure there is real transparency in the system, something that is currently lacking. We need to do this so that we can guarantee the continued success of our creative industries and fully benefit from what AI will bring. I want to make it clear, as others have, that these two sectors are not mutually exclusive; it is not a case of picking sides. Many in the creative industries are themselves users or developers of AI technology. We want to ensure that the benefits of this powerful new technology are shared, which was a point made by the noble Baroness, Lady Stowell, and her committee. It is obvious that these are complex issues. We know that the current situation is unsatisfactory in practice for the creative industries and the AI sector. That is why we have launched a detailed consultation on what package of measures can be developed to benefit both the creative industries and the AI sector. This is a genuine consultation. Many people from a range of sectors are engaging with us to share their views and evidence. It is important, and indeed essential, that we fully consider all responses provided in the consultation before we act. Not to do so would be a disservice to all those who are providing important input and would narrow our chance to get the right solution. I agree wholeheartedly with the noble Baroness and many other noble Lords, including the noble Lord, Lord Freyberg, on the importance of transparency about the creative content used to train AI. Transparency, about both inputs and outputs, is a key objective in the Government’s consultation on copyright and AI. This very ability to provide transparency is at the centre of what is required. The consultation also contains two other vital objectives alongside transparency: practical and clear control and reward for rights holders over the use of their work. This is quite the opposite of the notion of giving away their hard work or theft. It is about increasing their control and ensuring access to data for AI training. The Government certainly agree with the spirit of the amendments on transparency and web crawlers and the aims they are trying to achieve—that creators should have more clarity over which web crawlers can access their works and be able to block them if they wish, and that they should be able to know what has been used and by whom and have mechanisms to be appropriately reimbursed. However, it would be premature to commit to very specific solutions at this stage of the consideration of the consultation. We want to consider these issues more broadly than the amendments before us, which do not take into account the fact that web crawling is not the only way AI models are trained. We also want to ensure that any future measures are not disproportionate for small businesses and individuals. There is a risk that legislating in this way will not be flexible enough to keep pace with rapid developments in the AI sector or new web standards. A key purpose of our consultation is to ensure that we have the full benefit of views on how to approach these issues, so that any legislation will be future-proof and able to deliver concrete and sustainable benefits for the creators. The preferred option in the consultation is one proposal; this is a consultation to try to find the right answer and all the proposals will be considered on their merits. The Government are also committed to ensuring that rights holders have real control over how their works are used. At the moment, many feel powerless over the use of their works by AI models. Our consultation considers technological and other means that can help to ensure that creators’ wishes are respected in practice. We want to work with industry to develop simple and reliable ways to do this that meet agreed standards, in reference to the point made by the noble Viscount, Lord Camrose. Technical standards are an important part of this. There are technical standards that will be required to prevent web crawlers accessing certain datasets. Standards will be needed for control at the metadata level and for watermarking. I agree with the noble Viscount, Lord Camrose, that standards on the use of watermarks or metadata could have a number of benefits for those who wish to control or license the use of their content with AI. Standards on the use of web crawlers may also improve the ability of rights holders to prevent the use of their works against their wishes. We will actively support the development of new standards and the application of existing ones. We see this as a key part of what is needed. We do not intend to implement changes in this area until we are confident that they will work in practice and are easy to use. I also want to stress that our data mining proposals relate only to content that has been lawfully made available, so they will not apply to pirated copies. Existing copyright law will continue to apply to the outputs of AI models, as it does today. People will not be able to use AI as a cover for copyright piracy. With improved transparency and control over inputs, we expect that the likelihood of models generating infringing output will be greatly reduced.