On 6 January 2021, the AI Council published an AI roadmap which makes 16 recommendations to the UK government for UK AI strategy.

What is the AI Council?

The AI Council is a a non-statutory expert committee of independent members set up to provide advice to the UK government and high-level leadership of the AI ecosystem in the UK. It works to support the growth of AI in the UK, promote its adoption and use in businesses and society, and encourage experts to focus on priority topics in AI. 

AI Roadmap

The summary report by the AI Council sets out long-term ambitions and suggests near-term directions for all UK government departments, with the aim of cementing the UK as one of the best places in the world to live with, work with and develop AI.

The AI Council believes that the UK starts from a place of strength in research, enterprise and regulation. They urge the government to push for scale and reliability in areas of unique advantage and build a UK National AI strategy that scales up its investment for a decade and beyond.

Given the breadth and significance of potential AI applications, the Roadmap is concerned with overcoming the challenges of enabling change. It sets out suggested directions across three pillars: Research, Development & Innovation; Skills & Diversity, and Data, Infrastructure & Public Trust - and then, in the final section, addresses some specific measures to support adoption and the key areas of health, climate and defence.

The Council makes clear that the Roadmap is not a draft strategy, and nor is it an instruction manual. It is a set of recommendations and they recognise it will not be possible to achieve all at once, and this is why a National AI Strategy is needed to prioritise and set a time frame that will position the UK for success.

The AI Council's recommendations include: 

Research, Development & Innovation

  • Scale up public sector investment in AI; ensure consistent access to top global talent and find new ways to bring researchers, disciplines and sectors together. Build on the commitments in the government’s R&D Roadmap and suggestions in the soon to be published UKRI AI review.
  • Cement The Alan Turing Institute as a truly national institute, with a set of regional investments and provide assured long term public sector funding. (The Institute has published its response welcoming the recommendations.)
  •  Ensure R&D moonshots, as described in the R&D Roadmap, are both advancing and leveraging AI. (The Council suggests these could tackle fundamental AI challenges such as creating “explainable AI”.)

Skills and diversity

  • Commit to a programme of high level AI skill-building in the UK, including research fellowships, PhDs, Masters and apprenticeships.
  • Benchmark and track diversity levels to make data-led decisions about where to invest and ensure underrepresented groups are given equal opportunity.
  • Commit to achieving AI and data literacy for everyone. Create an Online Academy for understanding AI, with trusted materials and initiatives.

Data, Infrastructure and Public Trust

The recommendations in this section are the most interesting from a legal and regulatory perspective.

  • Lead development of data governance options and its uses. In particular, the Council notes that in the private sector, while regulators have begun good work to audit AI for data protection compliance, more work is needed to help businesses seeking to use data for AI by creating the conditions for the deployment of suitable privacy enhancing technologies. This should be furthered by accelerating work on translating the intent of a data sharing agreement into an actionable legal framework, and establishing guidelines for legal frameworks around different data sharing structures such as trusts, cooperatives and contracts.
  • Given that the detailed infrastructure needed to make AI work in different markets is often very context specific, create a National AI Strategy that links general principles to industry-specific arrangements of governance. One need shared by many markets is in establishing data infrastructure to permit companies to safely share and access data to develop common AI-related products. The public sector should lead the way with examples of good data sharing practices.
  • Clear and flexible regulation will be critical to supporting good data practices, and can help to level the playing field for new entrants into digital and data-driven markets. There remains much to be done in interpreting existing data rules to support research, development and innovation and well-resourced, independent regulators can assist in providing guidance.
  • A coalition of institutions and experts should work together to develop standard data sharing agreements compliant with data protection and intellectual property law .. providing practical exemplars, especially for business-to-business data sharing that could be adapted and adopted across the economy.
  • Ensure public trust through public scrutiny. The Council suggests that three key tenets: (1) clear transparency about automated decision making, (2) the right to give meaningful public input and (3) the ability to enforce sanctions could be encapsulated in a Public Interest Data Bill.
  • Commission an independent entity to provide recommendations on the next steps in the evolution of governance mechanisms, including impact and risk assessments, best practice principles, ethical processes and institutional mechanisms that will increase and sustain public trust.
  • Thoughtfully position the UK with respect to other major AI nations, building on existing partnerships.

National, Cross-sector Adoption

  • Increase buyer confidence and AI capability across all sectors and all sizes of company and support investment for local initiatives.
  • Support the UK’s AI startup vendor community, with greater access to data, infrastructure, skills, compute, specialist knowledge and funds.
  • Enable robust public sector investments in AI.
  • Use AI to meet the challenges of Net Zero carbon emissions.
  • Use AI to help keep the country safe and secure.
  • Build on the work of NHSX and others to lead the way in using AI to improve outcomes and create value in healthcare.

Between this report and the recent House of Lords AI report, the UK government now has a range of recommendations to consider from leading figures in the UK's AI sector.