UK government’s AI spend soars but data quality risks persist

UK government spending on IT and artificial intelligence projects has increased by 72% since 2018, with AI-specific contracts experiencing a rise of more than 1,000% over the same period. Despite this significant investment, experts are warning that the quality and accessibility of government data poses a significant risk to the nation's AI ambitions.

Spending trends

Government procurement for digital transformation projects reached GBP £18.9 billion in 2024, up from GBP £11 billion in 2018. The number of contracts awarded for AI-related work rose by 1,085% in that period, with a particularly sharp increase-64%-over the last year. The Department of Health and Social Care was the highest digital investor in 2024, spending GBP £1.45 billion, nearly GBP £450 million more than the next largest buyer, the Home Office, which spent GBP £1.01 billion. Other notable spending came from the Department for Transport, the Department for Work and Pensions, and Defra, with additional substantial contributions from the Metropolitan Police Service, Department for Education, and HM Treasury.

Data divide concerns

While headline funding figures imply a strong commitment to AI and digital innovation, a recent analysis by SAS and procurement specialist Tussell found that the practical effectiveness of these projects is hampered by what is termed a "government data divide." This divide refers to the gap between the high-quality, shareable data needed for successful AI deployment, and the siloed, inconsistent data that many departments continue to use.

Without improvements to data infrastructure and governance, experts believe public sector AI initiatives may remain limited in their impact. Fragmented and inconsistent data may restrict the scaling of projects and the realisation of efficiencies across government services.

"AI adoption in government is gaining real momentum. But our analysis shows that data quality and access remain the biggest barriers to scaling successfully - which puts the government at risk of a growing data divide.

To deliver AI at a national scale, data must be treated as critical digital infrastructure: planned, governed and maintained with the same rigour as physical assets like transport or energy. Without it, our report suggests, billions in AI investment risk becoming a collection of pilots rather than a platform for service-wide progress that will benefit UK citizens," said Dr Iain Brown, Head of AI & Data Science, SAS Northern Europe.

Infrastructure needs

The report recommends building stronger data-sharing frameworks, producing higher-quality digital records, and creating new datasets designed to be securely shared. Synthetic data, generated to mirror the statistical qualities of real information while protecting privacy, is identified as a practical step towards enabling safer AI training and collaboration between departments.

The analysis gives detailed insight into how digital capabilities differ across departments, suggesting that targeted improvements could yield the greatest benefit. While some departments have grown their digital investment rapidly, the variation in data readiness suggests there is still much work to be done for government services to leverage AI fully and consistently.

Calls for action

"Digital transformation is essential to reforming public services," says Gus Tugendhat, founder of Tussell. "The strong rise in IT procurement spend since 2018 shows that successive governments get this and are investing accordingly.

As this report demonstrates, however, a 'government data divide' is still an obstacle in the way of promoting much greater use of AI. Adoption of synthetic data would help to bridge this divide by enabling public sector teams to test, share and scale AI more safely. With clear AI governance and stronger data literacy, the UK government can seize the opportunities of AI to make public services work better for citizens."

Future direction

Building trust in government AI systems will require careful attention to data quality, sharing standards, and the secure development of digital infrastructure. Synthetic data is highlighted as a safe and effective means to accelerate these goals. Dr Iain Brown added: "As the government delivers on its AI and digital strategies, addressing the data divide will be critical to realising the UK's vision for trusted, citizen-focused AI. Synthetic data as part of the approach gives the government a practical, safe way to accelerate that progress and ensure that public trust in the technology grows alongside public-sector innovation."