To ask Her Majesty’s Government, further to the report by the Centre for Data Ethics and Innovation AI Barometer, published on 18 June, what assessment they have made of the benefits and risks of the use of artificial intelligence in addressing the impact of COVID-19.
My Lords, artificial intelligence played a very important role in responding to Covid, from identifying potential drug candidates to AI-driven education technology. AI also has the potential to drive productivity gains across sectors, supporting exciting new careers and businesses as an essential part of economic recovery. It is important that we keep society engaged as we do, so the centre’s Covid-19 repositories and its public attitudes surveys inform our understanding of public sentiment. The independent AI Council advises the Government on how best to realise the benefits and mitigate the risks.
I thank the Minister for her Answer, and I draw attention to my registered interest as a board member of the Centre for Data Ethics and Innovation. A year ago, the Prime Minister set out a vision, in his speech to the United Nations, for the UK to become a global leader in ethical and responsible technologies. We are discovering more deeply and painfully that ethics, good governance, human mediation and public trust are vital to realise the deeper benefits of these new technologies and prevent real harm. Will the noble Baroness affirm the importance of balancing innovation with a continued emphasis on ethics and good governance across the technology sector? In particular, will she confirm that the long-delayed government response to their own online harms consultation will be published this month, paving the way for much-needed legislation?
I thank the right reverend Prelate, all those involved in the Centre for Data Ethics and Innovation and all those involved in the ethics area for the important work they do. The UK remains absolutely committed to the ethical and humane deployment of AI and digital technologies, and it is absolutely right that we balance innovation with a continued emphasis on ethics and good governance. Further to the Prime Minister’s statement on this last year, we recently committed to global leadership on this issue as one of the founding members of the Global Partnership on AI, an international and multi-stakeholder initiative to guide the responsible development of AI grounded in human rights, inclusion, diversity, innovation and economic growth. On the second part of the question, we will be publishing our response to the online harms consultation shortly.
My Lords, will my noble friend welcome the establishment of the Global Partnership on AI earlier this summer? What are the Government’s hopes from it, and does she agree that we can make a success, nationally and globally, if we have human-led, innovative, inclusive and ethical AI?
My Lords, the AI Barometer talks of low levels of public trust being one fundamental barrier to the use of AI in both public and private sectors. Trust in government use of AI has been hugely damaged by the A-level algorithm fiasco. What are the Government doing to restore that trust? Is it not now crucial to put the CDEI on a statutory footing and ensure that there is a proper mechanism for ethical compliance across government services?
The noble Lord is absolutely right to focus on the importance of trust: it is a vital underpinning in the development of AI. I imagine he is aware that we have just published our National Data Strategy, which sets out very clearly the importance of public understanding of both government and non-government data within an ethical framework.
My Lords, back in April my noble friend Lord Darzi told the BBC in respect of Covid-19 that
“AI remains one of our strongest paths to achieve a perceptible solution but there is a fundamental need for high quality, large and clean data sets.”
Much of this data gets siloed in individual companies and universities, so what are the Government doing to unify these data sources to allow researchers to apply machine learning to new solutions for vaccines, monitoring and personalising care within an ethical context?
The noble Lord is absolutely right, although I am sure he would acknowledge that the quality, size and integration of those datasets vary considerably, as the recent report highlights, between different sectors of the economy. Again, the National Data Strategy and the consultation on it will be important mechanisms for addressing the issues that the noble Lord raised, as well as the open data initiatives and pilots that we are already running.
My Lords, the National Data Strategy has been published for consultation. Can the Minister assure us that the House of Lords will be included in that consultation? Perhaps the Minister would like to organise a webinar for interested Peers and guarantee that we can have a debate on the issues. The CDEI report notes that social care has been much less able to cope with providing data than the healthcare system because the level of training, funding and data collection in social care is so much lower. What plans do the Government have to help to improve that?
My understanding is that the consultation on the data strategy is open to everyone, but I am very happy to go back to the department and explore whether we can have a webinar for those Members of this House who are interested in taking part. Obviously, your Lordships’ Select Committee on Artificial Intelligence has been very influential already in our thinking. In relation to social care, the noble Lord referred to training and funding; it is also fair to say that the fragmentation of that sector is also a barrier to the adoption of AI, but we are also focusing on this.
My Lords, the EU White Paper of February 2020 promises legislation in Europe on public trust and regulation for AI next year. Will the Government commit to keeping our legislation on, and regulation of, AI in step with our former EU partners so as to benefit our creative industries?
My Lords, the endless development delays to the launch of the Covid-19 contact-tracing app, the Home Office decision, ahead of a judicial review, to scrap the controversial visa applications AI system, which was biased in favour of white applicants, and—most embarrassing of all—the A-level exam results scandal have all reinforced barriers to AI take-up as identified in this report, so what lessons have the Government learned from these fiascos?
I will make two points in response to the noble Lord’s remarks. First, it is the combination of the data and the human moderation that allows us to use AI most effectively. Secondly, the noble Lord highlighted some of the problems rather than some of the huge successes that we have had recently, including using AI and big data to identify those who needed to shield, using AI to predict the molecular structure of the virus, and many more important examples.
My Lords, the time allowed for this Question has elapsed.