Yesterday’s Spring Statement touched on technology, particularly defence and AI in public services. To truly position the UK as an AI superpower, we face a significant challenge: much of the government data necessary for developing cutting-edge language models isn’t ready for AI use. The “three pioneering DSIT-led Frontier AI Exemplars” announced require a focus on data quality and interoperability.
Data is crucial across every sector and underpins government operations, impacting every citizen. For the AI Opportunities Action Plan to work, we need to utilize data not just in our industrial strategy but throughout its implementation. We must view data infrastructure as a key growth sector, much like advanced manufacturing or life sciences. This requires a shift in how we manage and invest in our national data assets. Creating a National Data Infrastructure Roadmap as part of Invest 2035 is vital, something we pointed out in our response to the Invest 2035 green paper last year.
For real economic growth, we have to harness data insights to track progress and spot trends. A central dashboard aggregating data from government, industry, and regional partners would help policymakers see trends like new growth sectors or regional innovation clusters. Insights gathered can guide targeted interventions and investments, calling not just for existing data but also new data creation.
We can use advanced methods like network analysis to identify patterns in innovation systems. For example, progress in renewable energy tech could spur growth in specialized manufacturing or logistics, opening up fresh investment opportunities.
The government believes AI can free public sector workers from administrative tasks, allowing them to focus on vital services. The Alan Turing Institute estimates that civil servants handle 143 million complex tasks each year, and saving just one minute per task through AI automation could equate to about 1,200 person-years of work saved.
However, our latest research shows the UK government’s data isn’t AI-ready. We examined LLMs’ knowledge of government websites and statistics and found major flaws in data quality, accessibility, and interoperability, especially in the datasets available on data.gov.uk. These gaps, which previous governments have allowed to fester, could hinder the productivity gains that AI promises. The current administration has a chance to address this.
To make public and citizen data more FAIR (Findable, Accessible, Interoperable, and Reusable), we should adopt metadata standards like Croissant. This standard is designed for AI applications, allowing datasets on platforms like data.gov.uk to be indexed by tools such as Google Dataset Search, improving discoverability and usability. This approach will increase the readiness of government data for AI by enhancing interoperability and encouraging its widespread use by the AI community.
The government needs to capitalize on the potential of smart data schemes to spark innovation. Initiatives like Open Banking show how opening up data can benefit both businesses and consumers. By letting banks share data on products and transactions, we can create a secure and ethical flow of information across sectors. Imagine if we combined financial data with energy or transport data—it could lead to innovative services that drive economic activity.
To achieve this, we need cross-sector standards and governance. That means establishing a new central authority to develop these standards, ensuring interoperability and breaking down data silos. The UK should also engage with international efforts, like the Data Spaces in Europe, to align with global standards and promote cross-border data sharing.
Looking at data as a foundational layer of the economy highlights the need for a comprehensive 10-year roadmap supported by sustainable investment. A solid National Data Infrastructure Roadmap funded through multi-year commitments could provide the stability needed to achieve these ambitions.
Creating partnerships with leading organizations like the Alan Turing Institute, the Open Data Institute, and others will be crucial for developing interoperable systems, open standards, and AI-ready datasets. Working together will ensure the infrastructure fosters broad innovation and economic growth.
The government plans to establish a National Data Library for managing and accessing public sector data. If this platform is to provide secure and ethical access to public data, it must be designed to be AI-ready from the start. This includes implementing data hygiene measures endorsed by organizations like the ODI and establishing governance structures that balance innovation with public trust.
The UK’s data economy is already significant, growing from 6.5% of GDP in 2021 to 7.4% in 2023, a higher percentage than all European countries except Estonia. However, there’s still a lot of untapped potential. Only 21% of businesses that handle digital data analyze it for insights, and just 2% use it for AI or automated decision-making. Filling this gap is essential for ensuring responsible data sharing and unlocking economic value.
Recognizing data as a national asset can incentivize private investment and improve public sector data management. If we treat the data within the National Data Library as a valuable asset, we can secure the necessary funding for its development and maintenance. This will enhance data quality, accessibility, and innovation value.
Data is crucial for the UK’s economic growth and innovation and must be integral to our industrial strategy. Treating data as critical national infrastructure—like roads and energy—can empower the successful implementation of growth strategies across sectors. By unlocking public data, giving individuals control over their data, and strengthening organizational data capabilities, we can ensure data becomes a key player in the UK’s economic future. With the right investment and planning, we’ll realize the transformative potential of data, ensuring the UK remains a global leader in its use. However, we need to act quickly to keep up in the competitive digital landscape.
