“AI Roadmap”: UK AI Council calls for a national strategy

The coming months and years will be crucial in determining where the UK places its desired level of ambitions in AI. This is one of the key messages in the recent Roadmap published by the AI Council (“Roadmap”) in January 2021. The AI Council suggests that in order to position the UK for success a national strategy on artificial intelligence is necessary. As the Roadmap states, it is not a draft strategy and not an instruction manual. It contains a set of recommendations that reflect the view of the AI Council and 100+ additional experts and provides some insights into what the national AI strategy might look like.

What is the role of the AI Council?

The AI Council is a non-statutory expert committee comprised of independent members from either industry, the public sector or academia (such as Mastercard, the University of Cambridge, Microsoft Research Lab and The Alan Turing Institute to name a few). The intention is that members do not represent their organisations and not to promote their business with the committee; the intention is that they bring their expertise to contribute to the leadership of the AI ecosystem. Back in 2017, the review conducted by Professor Dame Wendy Hall and Jérôme Pesenti (“Hall-Pesenti Review”) recommended that the Government establish a UK AI Council to facilitate engagement between industry, academia, government and the public as AI in the UK will need to build trust and confidence in AI enabled complex systems. In late 2017, the Government announced that it was taking forward the recommendation in the Hall-Pesenti Review. However, the membership of the AI Council was only announced in May 2019. As the House of Lords’ recent report on AI notes (the “HL Report”) (which we have discussed in our article here), it is unclear why there was such a delay in getting the Council appointed. The purpose of the AI Council is to “put the UK at the forefront of artificial intelligence and data revolution”.

What might the AI national strategy look like according to the Roadmap?

The key underlying messages in the Roadmap are that the UK (i) needs to “double down” on recent investment it has made in AI; and (ii) must look to the horizon and be adaptable to change. The Roadmap sets out long-term ambitions and near-term directions for all government departments and recognises that it will not be possible to achieve them all at once. The Roadmap acknowledges that what can seem like a restatement of general principles that could be taken for granted is, in fact, a challenging environment for sustained systemic change. Broadly, the Roadmap suggests direction across four pillars:

1. support for research, development and innovation;
2. literacy in AI across the population with enhanced AI-related skills at all levels of educational and lifelong attainment;
3. sound physical, digital and virtual infrastructure, including robust and flexible regulation as a form of social infrastructure, resulting in good governance that boosts public trust; and
4. pragmatic approaches to national, cross-sector adoption that meaningfully advances public and private sector applications, with a focus on the key areas of health, climate and defence.

Research, development and innovation

The Roadmap states that the only way to realise AI’s full potential is to continue creating conditions for leading edge research. More specifically, it recommends to:

• scale up and make sustainable public sector investment in AI, ensure consistent access to top talent from around the world, find new ways to bring researchers, disciplines and sectors together, and build on the commitments in the Government’s R&D Roadmap published in July 2020 and suggestions in the soon to be published UKRI AI review;
• cement the Turing Institute as a truly national institute by providing assured long-term public sector funding that will give the Turing Institute and others the confidence to plan and invest for the medium term in strategic leadership for the UK;
• ensure that moonshots, described in the R&D Roadmap as challenge-led, high-risk, scalable programmes, are both advancing and leveraging AI. These could tackle fundamental challenges such as creating “explainable AI”, or important goals in any area where AI can contribute strongly, such as the UK Digital Twin programme or developing smart materials for energy storage in the move towards Net Zero carbon emissions. The Roadmap notes that properly designed and delivered moonshot programmes play to AI’s strengths by requiring people to work across boundaries and existing organisational structures. The specific vehicle for delivering moonshots requires further discussion.

In relation to retaining top talent, the HL Report made a similar recommendation and stated that the Government must make urgent steps to make clear to foreign researchers and students that they are very welcome and must ensure that any changes to the immigration rules promote rather than obstruct the study, research and development of AI.

Skills and Diversity

The Roadmap advocates for everyone to be able to live confidently with AI and suggests to:

• scale up and commit to an ongoing 10-year programme of high-level AI skill-building. This would include research fellowships, AI-relevant PhDs across disciplines, industry-led Masters’ and level 7 apprenticeships;
• make diversity and inclusion a priority by benchmarking and forensically tracking levels of diversity to make data-led decisions about where to invest and ensure that under-represented groups are given equal opportunities and included in all programmes; and
• commit to achieving AI and data literacy for everyone. An online academy for understanding AI, with trusted materials and initiatives, would support teachers, school students and lifelong learning.

The Roadmap states that since the AI Review in 2017, the Government has invested in schemes to promote graduate-level learning, which is a good start but not fit for the evolving landscape. Indeed, the HL Report referred to a recent survey by the Department for Business, Energy and Industrial Strategy that found that 28 per cent of people were positive about AI, while 20 per cent felt negative about it. A greater number of people said they were neither positive nor negative (44 per cent), with a further eight per cent saying they did not know.

Data Infrastructure and Public Trust

The Roadmap states that the UK should aspire to be the best place in the world to (i) access and use safe, secure and good quality data in order to develop new applications and business models; and (ii) exercise scrutiny of, and input to, automated decision making, in order to ensure the public can trust AI. To achieve this, the Government will need to put in place both physical and social infrastructure which will include broadband networks, high performance computing capacity, governance and good practice guidelines. In particular, the Roadmap proposes:

• to consolidate and accelerate the infrastructure needed to increase access to data for AI. It is worth noting that not all AI systems rely on access to the largest possible data sets. Indeed, some next generation AI technologies are designed to work with much smaller data sets, demonstrating that a tailored and context-specific approach will be required to facilitate optimal innovation and growth;
• for the UK to lead in developing appropriate standards to frame the future governance of data. Private sector regulators have begun the good work to audit AI for personal data protection compliance, but this should be further enhanced to create an actionable legal framework and guidelines for different data sharing structures, such as trusts, cooperatives and contracts. The Government will also play a crucial role in encouraging, or potentially requiring, AI companies to share research, analytics and data to promote competition and increase innovation;
• to ensure public trust through public scrutiny. Further exploration and scrutiny of appropriate governance mechanisms, particularly public sector adoption of AI, is long overdue. In addition, one approach of combining new forms of innovation and accountability is to involve the public in considering new ways to complement rights-based approaches such as consent; and
• to thoughtfully position the UK with respect to other major AI nations by building bilateral cooperation with countries like China and the US and strengthening existing partnerships with countries such as France and Canada. The Roadmap in particular cites the UK’s pivotal opportunity to play an influential role in shaping global AI adoption in its role as host of the G7 Summit this year.

National Cross-Sector Adoption

The Roadmap proposes the following recommendations:

• increase buyer confidence and AI capability across all sectors and all sizes of company. The UK needs to support investment for local initiatives to enable safe value-creating innovation and improve the data maturity needed for AI innovation;
• support the UK’s AI start-up vendor community, such as through initiatives like Tech Nation’s Applied AI Programme, enabling greater access to data, infrastructure, skills, specialist knowledge and funds. The Roadmap cautions against the struggles AI companies have remaining under UK ownership due to a current lack of government funding;
• enable robust public sector investments in AI, building capability in the use of data, analytics and AI to ensure intelligent procurement of AI as part of projects for public benefit. The Roadmap foresees, in time, the linking of public sector and private sector data, which will rapidly enhance the insights that can be gained through a more diverse data set;
• use AI to meet the challenges of Net Zero carbon emissions. We are already seeing AI contributing to a greater grasp of complex environmental and sustainability systems, from combatting illegal deforestation to understanding Arctic sea ice loss. The Roadmap urges the Government to consider how best to develop and showcase the UK’s ability in cleantech and its impact in moving towards Net Zero;
• use AI to help keep the country safe and secure, notably to detect and respond to increasingly sophisticated cyber-attacks. The Roadmap acknowledges that the ever-growing inter-connectivity of everyday life brings not only benefits but threats, and stresses the importance of engagement between government departments/agencies and defence and security companies to ensure AI is available to assess and respond to modern defence and security threats and opportunities; and
• build on the work of NHSX and others to lead the way in using AI to improve outcomes and create value in healthcare.

How is the UK positioned for global AI success?

The Roadmap suggests that the UK should leverage its founding membership in the Global Partnership for AI, a recommendation also endorsed by the HL Report. The Roadmap makes it clear that the scale of effort elsewhere must not be underestimated. For example, Germany has committed €3.1 billion towards a national AI strategy up to 2025 and France has pledged €1.5 billion up to 2022, with almost half allocated for research. The USA also recently announced investment of $1 billion over the coming five years to open 12 new institutes that will keep the country at the forefront of research in AI and quantum computing. In addition, as the HL Report notes, while according to the Global AI Index the UK is number one country for operating environment and third for research, it is only seventh for government strategy, eighth for infrastructure and eleventh for development. This reinforces the point that the UK must act swiftly on government strategy.

Next Steps

Based on the HL Report’s recommendations, it is important for the AI Council to concentrate on pragmatic issues, and to recommend to the Government the specific courses it should follow and actions it should take. The Liaison Committee of the House of Lords also thought that there must be increased and better coordination between various government and non-government departments and that a separate Cabinet Committee must be established to determine the strategic direction of Government AI policy and approve a five year strategy for AI. It remains to be seen how the Government will respond to the HL Report’s recommendation in respect of the Cabinet Committee. What is clear is that the AI Council will now take the next step and engage the wider AI ecosystem to discuss in a series of workshops how the AI community in the UK can collectively inform thinking on the Government’s future strategic direction.

We understand that he AI Council, together with the UK government's Office for AI, will arrange workshops in 2021 to engage on these topics. Interested parties are invited to participate in these workshops and can register to do so here.