Mr James Frith

I thank the hon. Lady for her intervention; I will expand on her point about transparency. We must have transparency, and it needs to be granular, enforceable and practical. AI developers must be required to disclose which copyrighted works they used to train or fine-tune their models. TollBit’s “State of the Bots” report confirms: “Whilst every AI developer with a published policy claims its crawlers respect the robots exclusion protocol, TollBit data finds that in many instances bots continue scraping despite explicit disallow requests for those user agents in publishers’ robots.txt files”. Many AI companies say that this need not hamper AI, and it is their voices that I wish to amplify today. This is about creating a fair, functioning market for training data that benefits all sectors. Last month’s YouGov survey of MPs and the general public agrees: 92% of MPs believe that AI companies should declare the data used to train their models, 85% say that using creative work without pay undermines intellectual property rights, and 79% support payments to creators whose work is used in training. The public expect us to do our best for our UK industries, and that is why we are square behind the Government’s instincts on British Steel. Let us apply the Government’s instinct here, too, as well as their strong record and rhetoric on digital images, deepfakes, online harms and the principle that if it is illegal offline, it is illegal online. Big tech always begins on the fringes before being regulated to the centre. We saw that most recently with age verification on app stores. Will the Minister commit to table Government amendments to the Data (Use and Access) Bill, in recognition of these supermassive concerns, to introduce a power to regulate for transparency, consent, copyright and compensation?