Online Safety

This Government are committed to keeping people safe online. For the first time, platforms now have a legal duty to ensure that they are protecting users from illegal content and, in particular, safeguarding children from harmful content, but we have gone further still. Within weeks this team have made self-harm and cyber-flashing, and now strangulation, extreme violence and pornography, priority offences. We will go further still by backing Ofcom to make sure that enforcement is robust too.

I thank the Minister for his answer, but the reality is that chatbots are prompting young people to commit suicide and to self-harm. What action can the Minister take to make sure that these chatbots are taken down and do not give this sort of advice?

I thank the hon. Member for raising these cases, which are very much in our minds. Each one is a deep tragedy. We have looked very carefully at this issue. Some chatbots, including live search and user-to-user engagement, are in scope of the Online Safety Act 2023, and we want to ensure that enforcement against them, where relevant, is robust. The Secretary of State has commissioned work to make sure that, if there are any gaps in the legislation, they will be looked at fully and robust action will be taken too.

I call the Chair of the Science, Innovation and Technology Committee.

The Minister says that the Government are looking deeply into this issue, but as part of my Committee’s inquiry into misinformation and algorithms, we heard conflicting evidence from Ministers and Ofcom as to whether generative artificial intelligence is covered by the Online Safety Act. The Government have refused to implement our call for legislation to bring generative AI under the same categorisation as other high-risk services. Under what circumstances is chatbot advice covered by the Online Safety Act, and will there be enforcement?

I thank my hon. Friend, both for the point she makes and for her ongoing insight and expertise on these questions. Let me be very clear about the current scope: chatbots that involve live search and user-to-user engagement are in scope of the Online Safety Act, as I mentioned. We are continuing to review its scope, and the Secretary of State has commissioned work. We will report its findings to the House.

I call the Liberal Democrat spokesperson.

I would like to join hon. Members in really pushing on questions about AI chatbots. Their human-like, assertive nature is filling a gap and many people, including children, are entrusting chatbots with medical opinions, legal advice and emotional support, with fatal consequences and without clear accountability. I know that this has been touched on, but it is really important. Ofcom explicitly includes only user to user or search, so one to one, which is actually where there are some of the most acute harms, is not covered. Will the Minister commit to working with Ofcom on classification, so we can ensure the responsible use of this technology and protect children from the unregulated harms of the growing dependence on chatbots?

Let me be very clear: of course we will. We have already both engaged with Ofcom and commissioned further work on this question, and we will report on that at the earliest opportunity.