Advanced Artificial Intelligence

Lord Holmes of Richmond Excerpts

Monday 24th July 2023 Lords Chamber

[Lord Holmes of Richmond (Con)]

My Lords, I thank the noble Lord, Lord Ravensdale, for securing this debate and congratulate him on the way he introduced it. I declare my technology interests as set out in the register.

When we come to consider the regulation of AI, it is first worth considering how we define it. There are multiple definitions out there, but when it comes to regulation, it is best not to draw that definition too tightly and perhaps better to concentrate on those outcomes that are intended and the challenges that we are seeking to avoid. Ultimately, AI is just the deployment of data, and it is our data, so a central pillar must be the explainability of how the AI comes to any decision, and how we should choose to regulate to achieve that level of explainability, which should be for the citizen understanding, not just from a software engineer’s perspective. Does the Minister feel that synthetic data offers a number of potential solutions, not least to the privacy questions, and what would the Government consider in terms of how they would go about the QA-ing and indeed the regulation of such synthetic data?

As has already been discussed, in that it is our data, it is right that we, and indeed every citizen, should have a say—should have a piece in this AI play. It will come down to trustworthiness, and everything that the developers, designers and businesses have to do to make this not trusted but trustworthy.

What more do the Government intend to do to have this level of public debate and discourse around such an existential issue? Similarly, does he agree that it would make sense to consider an AI officer on the board of all businesses of a certain size? I put an amendment to this effect down to the then Financial Services and Markets Bill, as AI is obviously already pervasive across our financial services industry. Would it not make sense for the Government to consult on having AI officers on the board of all businesses?

We have already heard a lot about ChatGPT—you cannot go a day without hearing about it—but what about the energy consumption that it took to train ChatGPT and for its continued use? Has my noble friend considered what the Government might wish to conclude on energy consumption of these AIs? Perhaps it would be better if photonic calculus was used, rather than more traditional math, to massively reduce the energy consumption of these systems.

Similarly, if the public are to be enabled, it will take much more than regulation. Does my noble friend agree that we should look at a complete transformation of our education system: data literacy, data competency, digital competency and financial AI literacy through every beat point of the curriculum. Would that not be a good thing for the Government to go out to consult on over the summer and the autumn?

If we are to make a success of AI—and it is in our human hands to do so—it will be only through the engagement and enablement of every citizen in every society, and understanding how to have that innovation in everybody’s human hands. If we stuff this up and it goes wrong, that will be not a failure of the AI or the technology but a human failure: of legislators, regulators, businesses, corporates and all of us.

What are the plans for the summit this autumn? How broadly will people be engaged? What will be the role for civil society at that summit? Finally, can my noble friend set out briefly what he sees as the key differences between the approach of the UK to AI and that of other jurisdictions, not least the European Union?

We can make a success of what we have in front of us, if we are rationally optimistic, understand the risks and step over the huge hype cycle of both unrealistic potential and overly described fears. We need to consider AI as incredibly powerful—but an incredibly powerful tool in our human hands, where we can grip it and make a success economically, socially and politically for all our citizens.