The AI Opportunities Action Plan is recommended reading

The 50 recommendations:

1. Set out, within six months, a long-term plan for the UK’s AI infrastructure needs, backed by a 10-year investment commitment. 

2. Expand the capacity of the AI Research Resource (AIRR) by at least 20 times by 2030 - starting within six months.

3. Strategically allocate sovereign compute by appointing mission-focused “AIRR programme directors” with significant autonomy.

4. Establish “AI Growth Zones” (AIGZs) to facilitate the accelerated build out of AI datacentres. 

5. Mitigate the sustainability and security risks of AI infrastructure, while positioning the UK to take advantage of opportunities to provide solutions. 

6. Agree international compute partnerships with like-minded countries to increase the types of compute capability available to researchers and catalyse research collaborations. 

7. Rapidly identify at least five high-impact public datasets to make available to AI researchers and innovators. 

8. Strategically shape what data is collected, rather than just making data available that already exists. 

9. Develop and publish guidelines and best practices for releasing open government datasets which can be used for AI, including on the development of effective data structures and data dissemination methods.

10. Couple compute allocation with access to proprietary data sets as part of an attractive offer to researchers and startups choosing to establish themselves in the UK and to unlock innovation.

11. Build public sector data collection infrastructure and finance the creation of new high-value datasets that meet public sector, academia and startup needs. 

12. Actively incentivise and reward researchers and industry to curate and unlock private datasets. 

13. Establish a copyright-cleared British media asset training data set, which can be licensed internationally at scale.

14. Accurately assess the size of the skills gap. 

15. Support higher education institutions to increase the numbers of AI graduates and teach industry-relevant skills. 

16. Increase the diversity of the talent pool. 

17. Expand education pathways into AI. 

18. Launch a flagship undergraduate and masters AI scholarship programme on the scale of Rhodes, Marshall, or Fulbright for students to study in the UK. 

19. Ensure a lifelong skills programme is ready for AI. 

20. Establish an internal headhunting capability on a par with top AI firms to bring a small number of elite individuals to the UK. 

21. Explore how the existing immigration system can be used to attract graduates from universities producing some of the world’s top AI talent. 

22. Expand the Turing AI Fellowship offer. 

23. Continue to support and grow the AI Safety Institute (AISI) to maintain and expand its research on model evaluations, foundational safety and societal resilience research. 

24. Reform the UK text and data mining regime so that it is at least as competitive as the EU. 

25. Commit to funding regulators to scale up their AI capabilities, some of which need urgent addressing. 

26. Ensure all sponsor departments include a focus on enabling safe AI innovation in their strategic guidance to regulators. 

27. Work with regulators to accelerate AI in priority sectors and implement pro-innovation initiatives like regulatory sandboxes. 

28. Require all regulators to publish annually how they have enabled innovation and growth driven by AI in their sector. 

29. Support the AI assurance ecosystem to increase trust and adoption.

30. Consider the broader institutional landscape and the full potential of the Alan Turing Institute.

31. Appointing an AI lead for each mission to help identify where AI could be a solution within the mission setting, considering the user needs from the outset.

32. A cross-government, technical horizon scanning and market intelligence capability which understands AI capabilities and use-cases.

33. Two-way partnerships with AI vendors and startups to anticipate future AI developments and signal public sector demand. 

34. Consistent use of a framework for how to source AI - whether to build in-house, buy, or run innovation challenges.

35. A rapid prototyping capability that can be drawn on for key projects where needed.

36. Specific support to hire external AI talent. 

37. A data-rich experimentation environment including a streamlined approach to accessing data sets, access to language models and necessary infrastructure like compute.

38. A faster, multi-stage gated and scaling AI procurement process that enables easy and quick access to small-scale funding for pilots and only layers bureaucratic controls as the investment size gets larger. 

39. A scaling service for successful pilots with senior support and central funding resource. 

40. Mission-focused national AI tenders to support rapid adoption across decentralised systems.

41. Development or procurement of a scalable AI tech stack that supports the use of specialist narrow and large language models for tens or hundreds of millions of citizen interactions across the UK.

42. Mandating infrastructure interoperability, code reusability and open sourcing. 

43. Procure smartly from the AI ecosystem as both its largest customer and as a market shaper. 

44. Use digital government infrastructure to create new opportunities for innovators. 

45. Publish best-practice guidance, results, case studies and open source solutions through a single “AI Knowledge Hub” accessible to technical and non-technical users.

46. In the next three months, the Digital Centre of Government should identify a series of quick wins to support the adoption of the “scan, pilot, scale” approach and enable public and private sector to reinforce each other.

47. Leverage the new Industrial Strategy.

48. Appoint AI Sector Champions in key industries like the life sciences, financial services and the creative industries to work with industry and government and develop AI adoption plans.

49. Drive AI adoption across the whole country. 

50. Create a new unit, UK Sovereign AI, with the power to partner with the private sector to deliver the clear mandate of maximising the UK’s stake in frontier AI.

Best foot forward. In a slightly odd change, the action plan first emerged with a foreword from the Prime Minister, only to be replaced on formal launch with one by the secretary of state for technology. Who can say whether AI was involved with this mischief - perhaps it was us hallucinating on our first read-through.

Finally, for now, Matt Clifford sums it up best as “business as usual is not an option,” emphasising that “government will need to be prepared to absorb some risk in the context of uncertainty”.

Does this plan and the full list of recommendations do enough to deliver the AI opportunity? What works well? What is missing? As always, I am keen to hear your responses to the plan and thoughts on what the government should be doing to regulate under uncertainty.