Much has been made in the media in the past few weeks of the Government’s AI action plan, unveiled at the end of January. Many have hailed the plan as a marker in the sand and welcome it as a forward thinking move towards supporting the advancement of the technology (not to mention positioning the UK as the ‘AI superpower’ of Starmer’s vision). Others have branded it a white elephant. Just last week, writing in The Times with an unsurprising lack of enthusiasm, leader of the opposition Kemi Badenoch described Labour’s general approach to AI as “a mess” which threatened to “stifle innovation” through overregulation. But which is it - white elephant or marker in the sand?
At MetadataWorks, we strive to help our customers use their data for good and naturally, we are in favour of any formal plans to support and invest in critical infrastructure that will put the UK at the forefront of the AI revolution - particularly when comes to investment into the underlying data resources that AI is reliant upon. That said, it’s a complicated topic with many nuances to consider.
Metadata management is essential for a self-sustaining AI cycle. To be effectively harvested, input, and processed by AI technologies, data must be findable, accessible, interoperable, and reusable (FAIR). In turn, AI algorithms should also follow FAIR principles to ensure maximum utility, transparency, and accountability.
According to recent research conducted by OneAdvanced, approximately one third of businesses have reported that their attempts to integrate artificial intelligence into their operations have failed. Clearly a robust strategy is needed in order to increase successful adoption and improve outcomes on any notable scale.
The plan in short
The UK government's AI Opportunities Action Plan outlines a strategy to position the UK as a global leader in artificial intelligence. Developed by tech entrepreneur Matt Clifford, the plan comprises 50 recommendations which can be broadly split into three topics; investing in AI Foundations, Promoting AI Adoption Across Sectors and Ensuring Homegrown AI Leadership.
I’m happy to see a good chunk of the plan looks at the foundations of AI. The document for example emphasizes the need to build secure and sustainable AI infrastructure, including a significant expansion of computational power, increasing government-owned AI computing capacity twentyfold by 2030 and constructing a new supercomputer to support academia and public services. Most notably, from a MetadataWorks point of view, the establishment of a National Data Library is proposed. This library aims to provide researchers and companies with access to valuable public sector datasets, facilitating advancements in various sectors – healthcare as a prime example. This is certainly an admirable aim, and an ideology that has been around for some time – that said, pragmatically it is no easy feat.
The National Data Library
The National Data Library (NDL) is, without a doubt, a cornerstone of the broader success plan outlined by the government. The official statement underscores this by declaring: “We will responsibly, securely, and ethically unlock the value of public sector data assets to support AI research and innovation through the creation of the National Data Library and the government’s wider data access policy.”
It’s encouraging to see significant engagement with the data and metadata community so far. However, while the vision is promising, there is currently a noticeable lack of detail regarding the implementation—particularly the ‘how.’ We anticipate more concrete information to emerge over the coming months.
For now, I'd like to offer my view on the Key Principles for NDL’s Success:
The EU’s progress with DCAT metadata profiles for descriptive data sets a strong precedent. The UK could benefit by aligning with these standards while establishing clear domain-specific terminologies and robust governance frameworks to improve and manage them going forward.
Conclusion
To fully realise the value of public sector data assets, the NDL must move beyond high-level ambitions and adopt a strategic, outcome-driven approach. This requires a strong foundation in interoperability, stakeholder collaboration, and measurable impact.
Metadata management is fundamental to this sustainable AI ecosystem, ensuring both data and algorithms follow FAIR principles—findable, accessible, interoperable, and reusable—to enhance utility, transparency, and accountability. This is exactly what we at MetadataWorks do—bridging the gap between data and impact, enabling our clients to unlock data’s full potential for the benefit of society.
By establishing clear standards and cultivating a culture of data excellence, the NDL can evolve into a transformative asset, accelerating AI research, driving innovation, and delivering real societal value.
What next?
In my view, the AI Opportunities Action Plan marks a positive start in terms of backing AI technology, however, more detail and further work is needed on standardising data across government research and linking to private sector data to make the plan a success. Further, I hope that data hubs such as ADR UK, SDR UK and HDR UK are considered by the powers that be to be fundamental to the success of the National Data Library and therefore the action plan as a whole.