2026 is the yr we should get critical about being a knowledge nation

The UK is starting 2026 with massive synthetic intelligence (AI) ambitions, however and not using a clear strategy to the information that underpins them. Regardless of some current constructive statements of intent from the federal government, knowledge wants extra consideration, significantly to make it AI-ready.

The final yr has seen consultations and techniques galore, together with the AI Alternatives Motion Plan, the 10-Yr Well being Plan for England, the Fashionable Industrial Technique and the Medicines and Healthcare merchandise Regulatory Company’s (MHRA) session on the Regulation of AI in Healthcare.

Many of those are both instantly associated to AI or rely closely on it to realize their goals. But the information on which their aim achievement depends is usually inconsistent, incomplete, or laborious to make use of at scale.

The newest – welcome – addition this yr is the Roadmap for Fashionable Digital Authorities, which interprets the sooner Blueprint for Fashionable Digital Authorities’s dedication to ‘utterly reshape public providers via digital transformation’ into particular departmental commitments and timelines operating to 2030.

The UK has robust knowledge foundations, together with a long-standing open knowledge custom and a number of the most useful public datasets anyplace on the earth, together with well being data, longitudinal schooling knowledge, official statistics, and cultural heritage knowledge.

The final Price range targeted on progress, tax modernisation and digital providers with the supply of virtually each measure linked to higher knowledge use – however a lot of the information that the federal government expects AI to depend on just isn’t prepared for that function. 

The ODI’s 2025 report, the UK authorities as a knowledge supplier for AI, revealed that enormous volumes of government-published knowledge can’t be reliably accessed by large-scale AI techniques resembling CommonCrawl, which is utilized by many main AI platforms, together with ChatGPT and Anthropic.

This might result in incomplete or deceptive outputs if these techniques are used to entry, for instance, advantages recommendation and well being info. It might in the end negatively impression folks’s lives in the event that they obtain inaccurate info, as a result of official sources can’t be learn by AI.

On the native stage, this might additionally have an effect on day-to-day service supply. Native authorities generate – and use – huge quantities of knowledge within the supply of social care, housing, transport and neighborhood providers.

There are alternatives for hard-pressed councils to make monetary financial savings by utilizing that knowledge because the uncooked materials for automation, forecasting and citizen-facing digital instruments. Nonetheless, our current report, Insights from UK councils on requirements, readiness and reform to modernise public knowledge for AI, printed collectively with Nortal, discovered that many councils are nonetheless publishing knowledge in codecs designed for reporting quite than superior technical use, which severely restricts the flexibility to make use of AI and digital instruments.

The challenges vary from inconsistent formatting and restricted metadata to inadequate infrastructure, together with lacking Utility Programming Interfaces (APIs), search performance, and model management.

Widespread failings got here via within the work, however there may be additionally an rising roadmap for progress. Our report means that “by modernising metadata, adopting requirements and enabling safe knowledge sharing, councils can transfer from ‘AI in concept’ to ‘AI that delivers on its promise’ – providers which might be environment friendly, clear and trusted.” Certainly, a number of councils are working to enhance knowledge high quality, strengthen infrastructure, and experiment with new approaches.

Constructing on these developments can be key to unlocking AI’s potential and making certain public datasets contribute extra absolutely to operational effectivity, service high quality and coverage effectiveness.

The institution of GDS Native in November 2025 to attach central and native authorities, and the Ministry of Housing, Communities and Native Authorities (MHCLG)’s completion of a discovery into GOV.UK One Login for native authorities (working with 50 native authorities), are constructive steps, however the structural challenges of inconsistent knowledge, restricted infrastructure, and provider focus stay vital.

Earlier this yr, the Division for Science, Innovation and Expertise (DSIT) printed its Tips and finest practices for making authorities datasets AI-ready, setting out high-level ideas, success standards, and examples of excellence to make sure that public sector datasets are AI-ready.

The rules cite fashions such because the ODI’s Framework for AI-Prepared Knowledge, which recommends concrete steps to assist organisations optimise knowledge for machine studying by enhancing its general high quality and adherence to requirements, and making certain its authorized compliance and accountable assortment.

Different key questions in regards to the public sector’s knowledge ecosystem must be answered within the yr forward.

For instance, regardless of £100 million in funding introduced for the Nationwide Knowledge Library in June 2025, and a dedication to share the programme’s imaginative and prescient and ‘kickstarter initiatives’ in early 2026, key questions are unresolved, together with whether or not it would function as public infrastructure or undertake a business mannequin through which sure classes of customers pays to entry the information it incorporates. This basic selection will form entry, licensing, and accountability.

The Knowledge (Use and Entry) Act units out frameworks for sensible knowledge and digital verification providers, however the Authorities nonetheless faces complicated points round AI and copyright. 

So there’s a possible paradox rising through which public knowledge is handled alternately as a public good and a monetisable asset: “sovereign” in concept however undefined in observe, with technical innovation shifting quicker than the guardrails wanted to make sure high quality, share worth, and defend our nationwide pursuits and particular person privateness.

That is clearest in well being and social mobility knowledge. The NHS’s well being knowledge is a priceless asset, but a lot of its worth has not been realised, partly due to folks’s comprehensible issues round privateness.

However there are methods that different kinds of – non-personal – knowledge can assist enhance well being outcomes and assist the three “shifts” set out within the 10 Yr Well being Plan (hospital to neighborhood, analogue to digital, and illness to prevention).

OpenActive, for instance, permits thousands and thousands of bodily exercise periods to be printed each month. This knowledge may very well be embedded in scientific pathways or social prescriptions and will energy providers that assist folks to get energetic, and be made accessible via the NHS App. Bodily inactivity has been calculated to price the UK over £20 billion a yr, so knowledge initiatives like OpenActive have the potential to each enhance well being outcomes and get monetary savings.

In social mobility, wealthy public datasets such because the Longitudinal Academic Outcomes dataset, the Guardian Pupil Match, and the PAYE dataset exist however aren’t coherently linked, leaving policymakers unable to reply fundamental questions, resembling why younger folks turn out to be NEET or how revenue and geography form alternatives for youngsters.

The issue just isn’t an absence of knowledge, however an absence of becoming a member of up round how this knowledge needs to be shared and used for public profit.

Most of the challenges are acknowledged in September’s Rewiring the state: Unlocking authorities transformation research. The Roadmap for Digital Authorities solutions them partly by setting out particular ambitions and commitments, together with the Division for Work and Pension’s (DWP) plan to modernise precedence legacy advantages providers, DEFRA’s migration of 19 crucial functions to the cloud by March 2026, and improved asset administration requirements this yr.

However questions stay about whether or not the tempo and funding will match the size of the problem.  The Roadmap for Digital Authorities undoubtedly strikes us nearer, however there may be extra to do. 

Knowledge is more and more framed as foundational nationwide infrastructure alongside energy grids and transport networks. The Roadmap commits to defining core knowledge requirements for interoperability and AI, and mandating API adoption throughout the general public sector by March 2026, however these foundations will solely succeed with sustained funding in assurance, stewardship, and the talents to place them into observe.

In the mean time, the UK underinvests in these quieter however important capabilities, and in consequence, its datasets aren’t but able to delivering the social and financial worth that they may.

If 2026 is to be a yr when digital and knowledge ambitions are realised, choices are wanted about which public datasets are nationwide infrastructure, which needs to be handled as public items, and the way worth will be realised – and shared – with out eroding public belief.

This additionally means investing within the important work of implementing knowledge requirements, governance, and digital and knowledge expertise to make sure AI techniques are constructed on stable foundations.

The Roadmap’s goal of 1 in ten civil servants working in digital, knowledge and cyber roles inside 5 years, and a core digital curriculum for all civil servants by April 2026, are welcome, however would require sustained dedication to make sure AI techniques are constructed on stable foundations.

The UK has the information property and the experience; what it must do now’s determine – intentionally and transparently – what sort of knowledge nation it intends to be.