Algorithms: Public Sector Decision-making - Question for Short Debate

Part of the debate – in the House of Lords at 7:55 pm on 12th February 2020.

Alert me about debates like this

Photo of Lord Wallace of Saltaire Lord Wallace of Saltaire Liberal Democrat Lords Spokesperson (Cabinet Office) 7:55 pm, 12th February 2020

My Lords, I first came into this whole area when I was a Lords Minister in the Cabinet Office seven years ago, when we were struggling with the beginnings of Whitehall going through the digital transformation. I am struck by just how much things have moved since then, mostly in a highly desirable direction, but we are all concerned that we continue to move with the right safeguards and regulations.

I am not an expert, but I have learned a lot from my son-in-law who is a financial quantitative analyst looking for when patterns do not hold as well as when they do, and from my son-in-law who is an systems biologist working on mutations in RNA and DNA, not that far away from the current Chinese virus. So I follow what the experts do without being an expert myself.

I am also struck by how very little the public are aware, and how little Parliament has been involved so far. The noble and learned Lord, Lord Keen, referred the other week to us returning to the “normal relationship” between Parliament and government, by which I think he meant Parliament shutting up more and allowing the Government to get on with things. I hope that is not what will happen in this area, because it is vital that the Government carry Parliament and then the public with them as they go forward.

A study for the Centre for Data Ethics and Innovation by MORI showed very little public awareness of what is going on in the sector. As the public learned, so they got more sceptical; I think the word used was “shocked”. We know that there are major benefits in the public sector from the greater use of artificial intelligence, if introduced with appropriate safeguards and regulations. This is evidence-based policy-making, which is what we are all interested in, so we need to make sure that we get it right and carry the public with the expansion of artificial intelligence.

There is a real danger of provoking a tabloid press campaign against the expansion of AI. We have seen what happened with the campaign against the MMR vaccine and how much credibility that got among the popular media, so transparency, regulation, education and explanation are important.

We need a clear legal framework. In 2012, one of the problems was that different Whitehall departments had different legal frameworks for how they used their data and how far they could share it with other departments. We need a flexible legal framework because, as we manage to do more things with mass data and mass data sharing, we shall need to adapt the framework—another reason why Parliament needs to be actively engaged.

We need ethics training for those in the public sector—and in the private sector interacting with the public sector—using artificial intelligence, so that they are aware of the limitations and potential biases and aware also that human interaction with the data and the algorithms is essential. One of the things that worries me at present, as an avid reader of Dominic Cummings’ blog, is the extent to which he believes that scientists and mathematicians should be allowed to get on with things without anthropologists, sociologists and others saying, “Hang on a minute. It’s not always as simple as you think. Humans often react in illogical ways, and that has to be built into your system.”

My noble friend Lord Stunell talked about public/private interaction. I think we understand that, while we are concentrating here on the proper public sector, one cannot disentangle private contractors and data managers from what goes on in the public sector, so we also need to extend regulation and education to the many bright private suppliers. I had a young man come to see me this afternoon who works for one of these small companies, and I was extremely impressed by how well he understood the issues.

We also need to engage civil society. Having spent a few weeks talking to university research centres, I am very impressed by how on top of this they are. There are some very impressive centres, which we also need to encourage. The richness of the developing expertise within the UK is something which the Government certainly need to encourage and lead.

My noble friend Lord Addington suggested that we may need to put the brakes on. We have to recognise that the pace of change is not going to slow, so we have to adapt and make sure that our regulatory framework adapts. I was pleased to listen to a talk by the director of the Centre for Data Ethics and Innovation hosted by the All-Party Parliamentary Group on Data Analytics last week. It is a very good innovation, but it needs to expand and to have a statutory framework. Is the Minister able to tell us what progress is being made in providing the CDEI with a statutory framework?

There are alternative approaches for the Government to take. One, the Dominic Cummings approach, would be to use speed and impatience in pushing innovation through and dismissing criticism. The second would be to go at all deliberate speed, with careful explanation, clear rules and maximum transparency, carrying Parliament and the public with it. The young man who came to see me this afternoon talked about having digital liberalism or digital authoritarianism—that is the choice.