– in the House of Lords at 1:34 pm on 26th April 2018.
My Lords, with the leave of the House, I will repeat a Statement made earlier today in the other place by my honourable friend the Minister of State for Digital and the Creative Industries. The Statement is as follows:
“The Government today publish our sector deal for artificial intelligence, a major collaboration with industry to secure the UK’s global leadership in artificial intelligence and data. From how we travel to how we live and work, AI holds transformative implications for every aspect of our lives and for every sector of the economy.
For the UK, the economic prize is clear, potentially adding 10% to our GDP by 2030 if adoption is widespread, with a productivity boost of up to 30%. In pursuing that prize, we start with strong foundations. The UK was recently ranked first among OECD countries in Oxford Insights’ AI government readiness index, and is home already to globally recognised AI companies including DeepMind, SwiftKey and Babylon Health. This success is supported by the UK’s strong combination of world-leading universities that drive skills and R&D; a thriving venture capital market for AI that leads among economies of comparable scale; and trusted universal public institutions such as our NHS that can pioneer data-driven innovation and connect the power of AI to the public good.
The sector deal that we have published today on GOV.UK therefore outlines how we are building on those foundations and on the independent review led by Professor Dame Wendy Hall and Jérôme Pesenti, reflecting that review’s spirit of partnership and consultation between government, industry and academia. In skills, we have made it the UK’s ambition to be home to the world’s best and brightest minds in artificial intelligence. We will support the Alan Turing Institute’s plans for expansion to become the national academic institute for AI and data science. We will create 200 additional PhDs in AI and related disciplines per annum by 2020-21, rising to at least 1,000 government-backed PhD places at any one time by 2025. We have set a target of 200 places for an industry-funded AI Masters programme, and will introduce an internationally competitive Turing Fellowship Programme in AI. We are also doubling the tier 1 exceptional talent visas to 2,000 a year to attract the brightest minds to the UK.
In infrastructure, we will ensure that the ambition of our AI sector is matched by the means of delivery in communications, in data and in supercomputer capacity. In telecommunications, we are investing over £1 billion to create a country with world-class digital capabilities, from 5G mobile networks to full-fibre broadband. In supercomputer capacity, we are delighted to announce as part of the sector deal that the University of Cambridge will make the UK’s fastest academic supercomputer, capable of solving the largest scientific and industrial challenges at breakneck speed, available to AI technology companies. That complements the Government’s support for start-ups’ access to hardware via the Digital Catapult’s Machine Intelligence Garage, and builds on Cambridge’s existing track record as a hub for AI and technology.
We are investing in data, too, because data is infrastructure. Just as roads help us to reach a destination, data helps us to reach a decision. For AI systems, data is the experience that they learn from to be able to process information and interact usefully with the world and the people who live there. This Government have always valued the economic benefits of pioneers having access to high-quality public datasets, but some of the most useful datasets for AI are those that organisations are reluctant to share with others—for instance, because they have commercial value. The world’s first centre for data ethics and innovation will therefore work to unlock the usefulness of that data while protecting its value for those organisations and, most importantly, keeping people’s data secure.
We want AI-led growth to be both empowering and inclusive, and that applies to our approach to data. But it also informs our commitment that the benefits of AI should be felt across the whole country. The sector deal makes a commitment to establish clusters and regional tech hubs, designed to power AI growth, across the entire UK. We will invest £21 million in Tech City UK over four years so that it can expand into Tech Nation, thus transforming the UK from a series of stand-alone tech hubs into a powerful network that can place the nation firmly at the top of global tech rankings.
The message is made clear by the investment that industry has brought. In total, its investment with government forms an investment package of nearly £1 billion. That support sits alongside the £250 million already allocated for connected and autonomous vehicles and the £1.7 billion that has been announced under the cross-sectoral industrial strategy challenge fund so far. But please be clear: our ambition in AI will not stop at this sector deal. This is only the start of UK plans to seize the opportunities of modern technology and ensure that it follows the highest ethical standards. By doing so, we will ensure we continue to build a Britain that is fit for the future”.
My Lords, I thank the Minister for repeating the Statement made earlier in the other place. We welcome it in its generality, although I have some comments about individual points. The most striking thing is that there is no reference in the Statement or the papers that accompany it to the excellent report recently published by your Lordships’ Select Committee on Artificial Intelligence. It may just be a case of the Government getting their retaliation in first. I hope not; I hope that in time they will respond positively and carefully to the various recommendations made by that excellent report and look forward to returning to the topic then.
AI clearly presents huge opportunities for the UK, and it is important that the Government are taking it seriously, as I think they are in the Statement. Responsibilities go with that initiative, as is evident in some of the points made, but I shall probe further on them.
On R&D spend, which is at the heart of the Statement, as reflected in the press statement issued jointly by DCMS and BEIS, the ambition is to get to 2.5% of GDP, with an eventual target, although it is not quantified, of more than 3%. Sweden, Austria, Germany, Belgium and Finland all have R&D expenditure of more than 2.5%, and South Korea spends more than 4% of GDP on R&D. If the UK really seeks to be a leader in AI and technology innovation, as we hope it does, why is the target so modest?
My second point is about the link back to young people and the school curriculum. It was suggested that the Government discouraged the independent review led by Professor Wendy Hall and Jérôme Pesenti from looking at the curriculum in primary and secondary schools, which you would think would be part of the process of trying to get our country as a whole geared up to do more across AI. If they were told that they could not go there because looking at the curriculum is very thorny and difficult, what on earth are the Government going to do about it? There is good news in the extra funding for teachers, but teachers do not create curricula; curricula have to be created in the wider context of education. I should be grateful if the Minister would comment on whether there is to be movement on that.
Also in education, there is the rather curious phrase that the Government are going to “create” 200 PhDs, as if they are something that you just print or issue, like coinage. Further reading and looking in more detail at other parts of the Statement should reveal that this will be funding for a welcome increase in the number of people taking PhD programmes. Presumably they will be independently offered by universities, not simply created by government diktat. However, are we not in the middle of a crisis of funding for higher education? Where in the Statement—I could not find it—is any reference to how the students will live on the additional PhDs that are being created? To narrow my question down, will the PhDs mentioned be part of the independent review of higher education, which is looking primarily at undergraduate courses but needs to look also at masters and PhD students?
We have looked at digital infrastructure time and again in this House, and each time the Minister has come to the Dispatch Box and talked about what progress has been made he has been met by a torrent of scepticism and concern that the reality is rather different from what the Government think. At the heart of this must be a commitment from the Government to get ahead of the rather unaspirational USO that they are about to introduce and go to fibre to the premises. FTTP broadband is the only way we can take the benefit of the technology, invest and get the returns that we will need as a country. We are so far behind the EU average on FTTP, which is 24% penetration. We are at about 2.7% penetration. Countries such as Portugal, Latvia and Lithuania have coverages of 86%, 85% and 81% respectively. What are the Government going to do about that? This will not get us to where we need to be.
On visas, there is a welcome suggestion that tier 1 numbers will be doubled, although that takes us to only 2,000—presumably per year. Will the Government reflect on whether that will be sufficient to reach the ambitions set out in the Statement?
I have two final points. In the Data Protection Bill, we have been concerned about whether sufficient resources and powers are available for the Information Commissioner to carry out her very responsible job of trying to ensure that we have a proper data regulatory structure. I understand that amendments are to be tabled that will increase the powers of the ICO, and look forward to discussing them when they reach the House—perhaps next week or the week after—but the question of resources is still open-ended. It seems that the Government will back and expand our AI activity. If that is the case, can they assure us that the additional resources required by the Information Commissioner’s Office will be provided at the appropriate time and that she will have the powers she needs?
Finally, on the very welcome news that the centre for data ethics and innovation is beginning to take shape and apparently has a budget of £9 million, what exactly is its current status? As I understand it, no legislative process has taken place, and I would be interested to know the timetable for that. Will the funding be limited to £9 million, or will other funds be available? More importantly, will it have a statutory position? The Government rightly pick up the need to ensure that all the work that is going on and is foreshadowed in the Statement will be effective for our economy, but it will be effective only if people trust that their data will not be abused and that there is appropriate understanding and a proper regulatory processes in place which engage with the ethical issues. We need a little more information on that. I should be grateful if the Minister could respond on when that will happen.
My Lords, having immersed myself in the subject of AI for the past year, I am absolutely clear that there is complete cross-party consensus on the potential for AI in the UK. I welcome today’s sector deal, particularly the evidence of cross-departmental working, which underlies quite a lot of the work that is beginning to take place. I very much hope that today’s sector deal is simply the tip of the iceberg of the Government’s AI policy and ambition. I note that the Minister used the word “ambition”, and I very much hope that this is but the first in a number of steps that need to be taken.
I hope we will have a much more extensive debate when the Government’s response to our Select Committee report is issued in due course, because it covers so many aspects. As I see it, today’s sector deal is essentially a nailing down of the commitments made in the industrial strategy, the proposals in the Hall-Pesenti review and the commitments made in the last Budget. I should be very interested if the Minister could unpack how much actual new money is involved in today’s sector deal, because I see it essentially as a packaging up for the sector rather than a new, dramatic development.
There are many aspects of the sector deal to welcome, not least the role of the British Business Bank in helping finance AI developers, growth companies, and so on. I hope they will be given an even more important role in the future, and I hope they will not go the way of the Green Investment Bank, which is an absolute object lesson for the Government in this respect.
The Select Committee thought that the fundamentals of government policy were right but it was a question of scale, ambition, co-ordination and drive behind the policies of the new bodies involved. There are many examples of this. The noble Lord, Lord Stevenson, rightly mentioned infrastructure investment. When only 3% of the country is covered by ultra-fast broadband, a £1 billion investment is neither here nor there. It is a bit of encouragement but it will not move us very fast up the curve compared to our international competitors. Then again, the scale of the skills gap is absolutely huge. I know that there was some negotiation as part of the Hall-Pesenti review, but 200 new PhDs in AI, as mentioned by the noble Lord, Lord Stevenson—off-the-shelf or not—being initially financed is the absolute bare minimum required.
Then again, we are heavily dependent on skilled EU workers. A Brexit brain drain is already threatening the UK tech sector, which relies heavily on foreign talent from the EU. DeepMind is already setting up a laboratory in Paris because of that. We need overseas students to stay. Will the Government reinstate post-study work visas for graduates in STEM subjects who find suitable employment within six months of graduating? The noble Lord, Lord Stevenson, mentioned a doubling of tier 1 visas. That is very welcome but why do not the Government declare, as the Select Committee suggested, a shortage occupation in tier 2 for machine learning and computer skills? That might make a huge difference. Collaborative research with EU countries is at risk as well. How will we fill the gap post 2020?
As virtually every Select Committee witness told us, creative skills will be crucial in the mix as well. What are the Government doing to emphasise not just STEM but STEAM in our schools? There is a dangerous dropping off of arts and creative subjects already. But, of course, it is not simply about the opportunities, of which there are many, but mitigating the risks as well, and making sure that we retain and build public trust in the new technologies involved. Inclusion is of crucial importance in this context. A strong inclusion and diversity agenda ran through our Select Committee report, which has been welcomed. In particular, we need more women in digital roles to help fill the skills gap. What are the Government doing to develop a culture that is inclusive, respectful and encourages women to pursue careers in AI?
Ethics must likewise be moved forward. I hope that the Government move forward quickly with this via the Centre for Data Ethics and Innovation by convening an international conference and other forms of international collaboration. I include the EU in this. Yesterday it published its report, Artificial Intelligence for Europe. In that, the role of the Charter of Fundamental Rights is highlighted as being the instrument by which one could incorporate a code of ethics. This makes the vote on Monday doubly valuable and I hope the Government will take due note. That is a very helpful way of making sure that we have an ethical framework that could cover most European countries.
I could raise many issues, not least data, which the noble Lord, Lord Stevenson, mentioned. I hope the Government will be talking to the Competition and Markets Authority about issues such as data monopolies. I hope that, as the Data Protection Bill goes through the Commons, they will look at whether we have real strength, and whether Article 22 of the GDPR really gives us sufficient rights of explainability for autonomous decision-making, as I raised in this House.
Finally, it is about ambition. If the UK wants to be seen as a world leader in any aspect of AI development, it needs to move as quickly as other countries, such as Canada and France. It must set its ambitions high to be a global player. It must welcome talent in growing its AI industry from start-ups to the next level.
My Lords, I am grateful for the many questions that I have to answer from the two noble Lords. I obviously should start by paying tribute to the committee of the noble Lord, Lord Clement-Jones. There was no reference to it in today’s Statement, and I take it as a compliment that the noble Lord, Lord Stevenson, thinks that DCMS works so quickly that we should include it in the sector deal a mere two or three weeks after it was published. I can say that we very much welcome the report. We thought it was a good piece of work and, in due course, we will provide a response. The report will help to inform actions going forward. It is important to understand that the sector deal today is only the beginning. When the noble Lord talks about the tip of the iceberg, that is very true. There are some things we intend to do, with facilities to make sure that they are monitored properly in the office of AI within the Government. I pay tribute to the noble Lord and his committee for that, and we will certainly look at that carefully.
Both noble Lords spoke of the skills gap. The noble Lord talked about Korea when referring to the 200 new PhDs, but we are not talking about North Korea; we are not just going to create 200 PhDs a year. They are proper PhDs that the Government will fund, leading to 1,000 government-funded extra PhDs by 2025. They are critical for the future but they are not the only areas in skills. The 200 have already been financed and there will be 450 by 2021 and 1,000 by 2025. They are starting in a phase-and-accelerating fashion in numbers per year.
Talking of skills and education, I accept, and have said before, that creativity is important. The Digital Catapult has identified the creative industries as one of the two high-profile potential areas for AI business growth in the UK. We understand that it is not simply a question of computer science, mathematics and such areas. To use the benefit of AI, we need creative minds. The businesses that already exist where we have a leading role in the world, have absolutely accepted that. One of the points of having the AI council is that it will bring together the Government, academia and the sectors to make sure that these points are raised at the highest level.
The noble Lord, Lord Stevenson, talked in particular about digital infrastructure and the commitment to fibre to the premises. We absolutely understand that we are behind many countries in fibre-optic connectivity. What he did not say is that we are ahead of Europe in superfast broadband by a long way, but we absolutely understand that we cannot be complacent. We are moving towards fibre to the premises. That is our goal and we absolutely accept that it needs to be done.
On visas, both noble Lords said that they welcomed the doubling of exceptional talent visas. They are for exceptionally talented people. We need to come to an understanding about the need for the new rules for immigration—luckily my noble friend from the Home Office is sitting here who will be very interested in this. The noble Lord, Lord Stevenson, talked about cross-government work on this, and the noble Lord, Lord Clement Jones, mentioned evidence. Our job is to make sure that the Home Office understands that when we come up with future Immigration Rules—we absolutely understand this is international business—we will need to have the best minds from around the world here. They will be attracted by our leading universities and the opportunities that will exist, and which this sector deal is trying to encourage.
The noble Lord, Lord Clement-Jones, talked about funding. When some of these things are mentioned, how much is actually new funding is a valid point. We have talked about just under £1 billion for this sector deal. Of this, about £600 million is new spending, and £342 million is existing spending that has either been repositioned or is in place already. Of that £600 million of new spending, about £300 million comes from the Government and, very encouragingly, £303 million from industry and the sector. For example, £35 million is from a Japanese venture capital company opening its first European HQ in the UK, £10 million is from Cambridge for the supercomputer, and there are others. About two-thirds is new money.
We absolutely accept that diversity is important, not only because it is the right thing to do, which it is, but because of all the talent we need to go forward. We have introduced the tech talent charter specifically to address that. Three weeks ago, I was at the G7 in Montreal talking about this and it resonated. In fact, we were held up in lights for it. We have 180 firms signed up and aim to have 500 by the end of the year. It is meaningful, and not just motherhood and apple pie about what we wish to do, because one of the things that firms sign up to is providing data centrally on the diversity aspects of their business so that we can compare and see that there is actual and meaningful progress. The charter will give organisations tangible actions and principles that they can adopt to become more gender-diverse.
I think that answers most of the questions. I am grateful for the broad welcome that both noble Lords have given.
My Lords, the world of artificial intelligence is advancing rapidly and changing the whole time. Is my noble friend satisfied that our laws are up to date and can cope with the advances being made?
My noble friend has the advantage of having been on the committee and probably knows more about this than I do. I do not think that one could ever say that one was satisfied that the laws were perfect in a fast-moving field such as AI and the new tech area. The Data Protection Bill, which is coming up for Report in the other place soon, is one way in which Europe and this country are bringing in data protection. In that context, I should mention the Information Commissioner, referred to by the noble Lords, Lord Stevenson and Lord Clement-Jones. We are looking carefully at what the Information Commissioner has asked for, especially in terms of powers. We are working on the legislation and trying to make it as future-proof as possible. Whether the Data Protection Act will last the 20 years that the last one did, I am not so sure.
Much of the outcome of all the work, which I very much welcome, is intangible. Who is going to own this intangible property? It is all right when it is used for the public good, but what happens when it is used for private profit? Surely this is the basis of the dispute over the work of Cambridge Analytica, and has to be settled before we put a lot of money into developing all this intangible property.
That is exactly why we are setting up the centre for data ethics and innovation. It will be a world-leading institution. Artificial intelligence is a force for good and potentially a force for evil. We absolutely acknowledge what the noble Lord says, but we are specifically addressing that. I was also asked about the timetable for the centre. The chair is being recruited now and we hope to have it up and running by the end of this year. It will have a statutory basis in due course, but will be up and running before then because, as the noble Lord rightly says, we have to address some of these problems. For example, the report talked about data trusts, to make sure that public and private data are available in a sustainable way and benefit SMEs as well as the enormous organisations.
My Lords, the Germans will make smart cars; the United States and Canada will focus on the internet. We have a real opportunity in the United Kingdom to do ethical AI, not least when we consider the areas of finance, law, research and biotech. Does my noble friend agree that when it comes to AI in the UK, the only way is ethics?
The only way to live your life is in ethics—not in Essex. As far as this is concerned, it is also important to collaborate internationally. The Prime Minister announced a new partnership with the World Economic Forum at Davos on developing a framework for the responsible procurement of AI in the public sector. That is one example of how we need to work with other organisations. We will continue to work with the EU while we remain a member and hope to negotiate a sensible arrangement on exiting for exactly that reason.
Have these matters been devolved to the various Administrations in Scotland, Wales and Northern Ireland and, if not, what discussions do the Government have in mind on these most impactive matters?
As I have said, issues such as AI and data have to be dealt with internationally as well as in the UK. The sector deal includes the devolved Administrations. My department continues to have regular discussions about these issues with the devolved Administrations.
My Lords, the Select Committee report and the Made Smarter review made quite a lot of the opportunity of bringing smaller and medium companies into the AI revolution and using it to make themselves more productive and competitive. How does the sector deal help that?
One example that I think I have already mentioned is data trusts. The review made the point that big companies have not a monopoly, but the advantage of having so much data. SMEs and small companies need access to that data in order to grow. That is the whole point of AI. If we can get a mechanism that allows big and small companies to work together on datasets to retain the value and to get some use of it, it would be a great advantage. We are committed to having pilots on data trusts in place by the end of this year.
My Lords, I apologise for not having been here at the beginning of the Statement. My question relates to a narrow field: the issue of fully autonomous weapons systems which are using AI and learn as they go on. What is the Government’s position on the development of fully autonomous weapons systems, bearing in mind that we know that at least two countries are working on what I think is an extremely dangerous thing?
The development of weapons generally is a very dangerous thing. We consider that the existing provisions of international humanitarian law are sufficient to regulate the use of weapons systems which might be developed in the future as they have been flexible enough in the past to cope with the invention of new means of warfare such as submarines and aeroplanes, but we are obliged to determine whether new weapons or means comply with international law. We will continue to engage with the UN on this point. We bear it in mind; we understand the implications of it, and we will remain within international law as it stands.
My Lords, my noble friend stole my thunder a bit. In the way in which AI is described here, it sounds very benign. It is indeed important to innovation in the future, but it is stuffed with risks and dangers wherever you look, from labour markets to weaponry and all sorts of other areas. It is a huge mix of advantages and massive problems. I would like at least some comment on how the Government will deal with them.
The Statement repeated the idea that AI will inevitably increase productivity. I know where the statistics come from. I am deeply sceptical about them. The advance of the digital revolution so far has been associated with declining, rather than increasing, productivity. We have to be careful not to see some magic in all this which may not be there, which would then bring us back to the problems and dangers.
The Statement said that AI had the potential to bring about a massive increase in productivity. In some areas, it will, as case studies show. For example, KLM doubled the number of text-based customer inquiries it handled during the past year while increasing the number of agents by 6%, so it is possible. I understand that there will be disruption in jobs because there will be probably be an increase in the number of high-value jobs. It will have implications. Overall, we think that it has the potential to raise productivity if it is handled properly, and by quite a lot. However, we accept that it has problems. We have to encourage such things as lifetime learning to enable people to transfer their skills so that they can contribute in a more modern way.
We accept that there are problems and dangers. That is one reason why we will have the centre for data ethics and innovation: so that we can bring in independent people to advise the Government on where regulation will be necessary and how regulations and laws should be developed. We are addressing that. The AI council will also inform government, because it will not just be government mandating from the centre; it will be a place where academia, the sector, industry and government can come together to drive the changes in the future.