The appointment of tech entrepreneur and Chair of the Advanced Research and Invention Agency, Matt Clifford, as the Prime Minister's AI adviser suggests a pivot. Within days of winning the election, the government commissioned Clifford to prepare a report on how the UK could better harness AI.
Clifford's wide-ranging AI Opportunities Action Plan (the "Plan") consists of no fewer than fifty recommendations on how the government can 'turbocharge' the UK and drive change and economic growth through leveraging the power of AI. Clifford's recommendations point to a more structured and ambitious framework for AI development and deployment in the UK. The government's acceptance of these recommendations commits not only to fostering innovation, but also addressing the ethical, legal, and societal implications of AI.
All this was given fresh 'context' this week with the emergence of the Chinese-developed large language model ("LLM") 'DeepSeek', which rocked the global tech community (and chipmaker Nvidia's share price) by apparently rivalling the capabilities of its competitors at a fraction of the infrastructure and development costs. The UK will need to run hard to remain competitive in this fast moving sector.
This article summarises the main points from the Plan.
Dissecting the AI Opportunities Action Plan
The Plan is structured around three core pillars, each designed to position the UK as a global leader in AI:
Pillar 1. Laying the foundations to enable AI: this pillar focuses upon building the necessary infrastructure, data frameworks and talent pipelines to support AI innovation and deployment.
Pillar 2. Encouraging the cross-economy adoption of AI: this pillar emphasises the integration of AI technologies across various sectors to improve public services, drive economic growth and enhance quality of life.
Pillar 3. Securing our future with home grown AI: the final pillar aims to develop sovereign AI capabilities, ensuring that the UK maintains control over critical technologies and can influence global AI standards and governance.
These pillars collectively outline the UK's strategic approach to harnessing AI's potential whist being alive to certain associated challenges. We consider some key points further below.
i) The role of data and the National Data Library
Data is the lifeblood on which LLMs depend, being essential to their 'training'. The Plan encourages the government to 'unlock' the valuable datasets that it holds, suggesting that this will encourage innovation from UK start-ups and researchers, whilst simultaneously attracting talent to the UK. The Plan also champions the planned establishment of the National Data Library, an initiative aimed at centralising and standardising access to valuable public sector datasets, with the purpose of facilitating research and development across various sectors.
Whilst such datasets will undoubtedly be an invaluable asset to the AI revolution, it is vital that the government (and institutional users of LLMs) find a balance between innovation and safety/data protection, ensuring that they supplement the drive to innovate with effective deployment and security measures. This particularly is an area requiring considerable upskilling given the current state of play in many sectors. The changes to the data protection regime proposed in the Data Use and Access Bill and regulatory comment from the government since then suggest the degree of mandated control and regulation is likely to be lower than the EU, and hence appealing to AI vendors.
ii) The 'Scan-Pilot-Scale' approach and public procurement
To integrate AI into public services effectively, the Plan introduces the 'Scan-Pilot-Scale' methodology, which encourages public sector organisations to:
- Scan for potential AI applications.
- Pilot projects to assess feasibility.
- Scale successful initiatives across the board.
This strategy necessitates a re-evaluation of public procurement processes to accommodate the iterative nature of AI development. Traditional procurement models, which encourage fair competition and transparency, may need to be adapted to allow for greater flexibility and collaboration with AI vendors.
iii) Updating copyright laws
Pointing to the European Union's AI-forward approach to intellectual property rights, the Plan calls for the government to implement a competitive copyright regime in favour of AI innovation.
Given the complex challenges posed by the intersection of AI and intellectual property, there is little surprise that this was one of the areas with which the government only 'partially agreed'.
To help inform its next steps, the government has initiated a consultation on copyright and Artificial Intelligence to explore how existing laws can be updated to reflect the realities of AI-generated content and the use of copyrighted materials in training AI models. This consultation seeks to balance the rights of creators with the need for accessible data to fuel AI advancements. Given the significant copyright infringement litigation ongoing in the United States in relation to LLMs, it is vital that the government strikes the correct balance of supporting the AI sector without abrogating private law rights (or alienating their holders).
iv) Progressing towards AI regulation
In line with global trends, the UK acknowledges the necessity of regulating AI to mitigate risks and ensure ethical deployment. While silent on the need for regulatory change, the Plan does stress the importance of regulators acting in favour of safe AI innovation, but not if it "blocks the path towards AI's transformative potential". The Plan suggests that the government may wish to consider creating a central body with powers to overrule regulators on matters pertaining to AI. If implemented, it will be interesting to see how a hierarchy of regulators and their competencies emerge. The government must beware the obvious risk of unintended consequences, and the degree of risk it will accept for the public's data to leverage the transformative potential.
v) Creation of UK Sovereign AI
The Plan advocates the development of UK Sovereign AI capabilities, aiming to reduce dependence on foreign technologies and enhance national security. This initiative mirrors efforts in other nations, such as China's emphasis on indigenous innovation and the European Union's push for technological sovereignty.
By fostering home-grown AI solutions, the UK intends to bolster its strategic autonomy in the digital age and become an 'AI maker' rather than an 'AI taker'. However, LLMs are notoriously greedy for energy, water, and physical space, and the UK's potential for 'AI independence' could collide abruptly with other government initiatives, not least Net Zero – though areas with excess power generation such as coastal Scotland are mentioned.
Practical guidance
The Plan presents both opportunities and challenges for businesses operating within and beyond the UK, we consider some of the key opportunities and challenges below:
Engaging with funding and public procurement
Those interested in securing funding for AI research, development and innovation should monitor announcements from government bodies regarding grants and investment programmes. Those interested in building and developing AI infrastructure should similarly look out for funding opportunities. Also, understanding the 'Scan-Pilot-Scale' approach will be crucial for those looking to participate in public sector projects, as procurement processes may evolve to prioritise flexibility and collaboration. Once the National Data Library is live, innovators and researchers should consider engaging with it, which will provide access to valuable datasets for research and development to kick-start AI innovation.
Navigating international operations
For businesses operating across multiple jurisdictions, particularly in the European Union, it is essential to recognise the increasingly divergent regulatory landscapes. The EU's stringent AI Act contains more rigorous compliance requirements compared to the UK's flexible approach to date, which suggests a more adaptive framework – which the Plan confirms. Staying informed about these differences and seeking legal resource to help navigate them will be vital to ensure compliance and mitigate risks.
DWF comment
The Plan marks a significant evolution in the UK's approach to AI. By addressing critical areas such as data accessibility, public sector integration, intellectual property, regulation, and sovereign capabilities, the Plan lays a robust foundation for future growth. The success of the Plan will depend upon practical considerations, including funding allocation, infrastructure development, and international collaboration.
We will be providing further insight and commentary on aspects of the Plan, including: Energy, AI/Data Protection, Intellectual Property Rights, Competition Law and Public Procurement.
If you have any questions in relation to AI, please get in touch with one of our team below.
Written with Maniva Bi and Victoria Anning.
