Contribution
My Lords, the technological capabilities and their misuse that have prompted this Statement are, needless to say, deeply disturbing and demand our careful attention. The use of AI to generate non-consensual sexual imagery of women and children is both grotesque in itself, but also corrosive of trust in technology more broadly.
We therefore welcome the Secretary of State’s confirmation that new offences criminalising the creation or solicitation of such material will be brought into force this week. We support the enforcement of these laws. We also welcome Ofcom’s decision to open a formal investigation into the use of Grok on X under the Online Safety Act, an investigation that must proceed swiftly to protect victims and hold platforms to account.
Hard though it is to predict the misuses of emerging technologies, we must collectively find better ways to be ready for them before they strike. I fear there is a pervasive and damaging sense of regulatory, legislative and political uncertainty around AI. As long as that remains the case, we risk remaining a victim of events beyond our control.
From the outset of this Parliament, and indeed in opposition, the Government have pledged to legislate on AI. Reviews and policy documents, including the Clifford AI Opportunities Action Plan, promised a framework to drive adoption and regulatory clarity. However, we still have no clear timeline, nor even a clear account of the Government’s policy on AI.
It is worth noting that the legislative tools the Government are now relying on to implement their proposed new offences, such as the creation and solicitation of non-consensual intimate images, are the product of amendments introduced by this House to the Data (Use and Access) Act. Ministers have repeatedly argued both that binding AI regulation must come, and that the existing multi-regulator framework is sufficient.
Evidence to the House of Commons Science, Innovation and Technology Committee late last year confirmed that the Secretary of State would not commit to a specific AI Bill, instead speaking of considering targeted interventions rather than an overarching legislative framework. This may indeed be the right approach, but its unclear presentation and communication drive uncertainty that undermines confidence for investors, businesses and regulators, but above all for citizens.
Progress on other AI-related policy commitments seems to have stalled too. I do not underestimate the difficulty of the problem, but work thus far on AI and copyright has been pretty disappointing. I am not seeking to go into that debate now, but only to make the point that it contributes to a widespread sense of uncertainty about tech in general and AI in particular.
Frankly, this uncertainty has been compounded by inconsistent political messaging. Over the weekend, reports emerged that the Government were considering banning X altogether before subsequently softening that position, creating wholly unnecessary confusion. At the same time, the Government have mischaracterised X’s decision to move its nudification tools behind a paywall as a means to boost profits, when the platform argues, reasonably persuasively, that this is a measure to ensure that those misusing the tools cannot do so anonymously.
Nor has there been much effective communication from the Government about their regulatory intentions for AI. This leaves the public and businesses unclear on how AI will be regulated and what standards companies are expected to meet. Political and legislative uncertainty in this case is having real consequences. It weakens our ability to deter misuse of AI technologies; it undermines public confidence, and it leaves regulators and enforcement agencies in a reactive posture rather than being empowered to act with a clear statutory direction.
We of course support efforts to criminalise harmful uses of AI. However, under the Government’s current Sentencing Bill, most individuals convicted of these new AI-related offences against women and girls will be liable for only suspended sentences, meaning that they could leave court free to continue using the technology that enabled their crime. This is concerning. It cannot be right that someone found guilty of producing non-consensual sexual imagery may walk free, unrestrained and with unimpeded access to the tools that facilitated their offending.
As I say, we support Ofcom’s work and the use of existing powers, but law without enforcement backed by a coherent, predictable regulatory regime will offer little real protection. Without proper sentencing, regulatory certainty and clear legislative direction for AI, these laws will not provide the protection that we need.
We urge the Government to publish a clear statement on their intentions on comprehensive AI regulation, perhaps building on the AI White Paper that we produced in government, to provide clarity for both tech companies and the public, and to underpin the safe adoption of AI across the economy and society. We must assume that new ways to abuse AI are being developed as we speak. Either we have principled, strategic approaches to deal with them, or we end up lurching from one crisis to the next.