My Lords, as ever, the noble Baroness, Lady Kidron, has nailed this issue. She has campaigned tirelessly in the field of child sexual abuse and has identified a major loophole.
What has been so important is learning from experience and seeing how these new generative AI models, which we have all been having to come to terms with them for the past 18 months, are so powerful in the hands of ordinary people who want to cause harm and sexual abuse. The important thing is that, under existing legislation, there are of course a number of provisions relating to creating deepfake child pornography, the circulation of pornographic deepfakes and so on. However, as the noble Baroness said, what the legislation does not do is go upstream to the AI system—the AI model itself—to make sure that those who develop those models are caught as well. That is what a lot of the discussion around deepfakes is about at the moment—it is, I would say, the most pressing issue—but it is also
about trying to nail those AI system owners and users at the very outset, not waiting until something is circulated or, indeed, created in the first place. We need to get right up there at the outset.
I very much support what the noble Baroness said; I will reserve any other remarks for the next group of amendments.