Lord Young of Acton (Con)

My Lords, I oppose government Amendment 429B in this group. I declare my interest as the director of the Free Speech Union. Like my noble friends, I will try to be brief. As several noble Lords have already pointed out, this amendment would grant the Secretary of State at the Department for Science, Innovation and Technology sweeping Henry VIII powers at a very late stage in our consideration of the Bill, thus giving this House far too little time to scrutinise them. Subsection (1) of proposed new Section 216A would grant the Secretary of State the power to “by regulations amend any provision of this Act”— the Online Safety Act— “for or in connection with the purposes of minimising or mitigating the risks of harm to individuals in the United Kingdom presented by” among other things, “illegal AI-generated content”. That will presumably include content that breaches Section 127 of the Communications Act 2003, meaning that it is grossly offensive. This can include memes. In 2024, a man called Lee Dunn was sentenced to eight weeks in jail for reposting three “grossly offensive” memes on Facebook, having pleaded guilty to a Section 127 offence. How will Ofcom monitor whether AI chatbots are generating grossly offensive content? Will the Secretary of State use the powers granted to her by this amendment to insist that spyware is installed on personal computers and mobile phones? Perhaps your Lordships consider that too remote a risk, but what about requiring technology companies to carry out client-side scanning of people interacting with AI chatbots on their devices—much like how Section 121(1) of the Online Safety Act grants Ofcom the power to require companies, including those that own private messaging apps such as WhatsApp, to scan content on people’s personal devices and report certain categories of illegal material to the National Crime Agency? Do not forget that this amendment would allow the Secretary of State to amend “any provision” of the Online Safety Act in order to minimise or mitigate the risks of harm posed by illegal AI-generated content. I dwell on this to illustrate just how wide-ranging and open-ended are the powers that this amendment would grant to the Secretary of State—powers that could have far-reaching consequences for civil liberties and freedom of speech. Another risk is the definitions part of the Amendment. Subsection (17) disapplies Section 59(14)(a) of the Online Safety Act when it comes to illegal AI-generated content. Section 59(14)(a) qualifies the scope of illegal content in Part 3 of the Act, and disapplying it gives the Secretary of State enormous scope to enlarge the definition of illegal content and impose proactive suppression duties on AI chatbots to make sure they comply with the new draconian censorship regime. If the Government believe there are specific harms that users of AI chatbots are currently exposed to and should be protected from—and I certainly do not say that there are not—let them bring forward primary legislation so we can consider the remedies they propose and factor in the trade-offs, particularly when it comes to free speech.