Max Wilkinson

The hon. Gentleman is entirely right. I agree that the concepts of opt-out and opt-in need to be pursued at greater length. Thirdly, the expert told me that the preservation of value offered stronger protection: “By requiring explicit permission, an opt-in model helps maintain the value of original creative works.” The expert did point out two drawbacks. First, an opt-in approach has potential drawbacks in the form of an extra administrative burden on creators. Interestingly, this expert’s second listed drawback was that an opt-in model would place limits on AI’s ability to gather data for training and development, which does not seem to me like much of a drawback for creators. I asked that very same expert what would happen if creators lost their intellectual property rights to AI. The expert told me there was a risk of “a loss of income and motivation, a devaluation of creative work, ethical concerns, legal uncertainty” and, intriguingly, “domination by AI operators.” I use the word “intriguingly” because this expert seems aware of its own power—the expert was Google Gemini. At this stage, those considerations are unknowns, and there is much uncertainty. Google Gemini is pulling information produced mostly thanks to human endeavour and discussion sourced from across the internet, but the fact that this view is being presented by AI itself surely suggests there is cause for some concern. Our role as parliamentarians must be to protect the interests of humans, not big tech companies; to scrutinise the proposals of big tech companies; to avoid the luddite tendency, crucially; and to build in suitable safeguards.