UK project tests how legal data can support AI use in government
The project findings highlight governance and data quality as needed foundations for trusted AI.
The UK Government Digital Service has highlighted data maturity as a key requirement for preparing public sector data for AI use.
The findings come from a project conducted with The National Archives, part of GDS’s wider work to ensure public sector data is managed as a strategic national asset.
During a discovery phase completed in April 2026, the organisations assessed whether legal data, including legislation and case law, could be prepared for AI applications. The work focused on governance, data quality, organisational readiness, and the risks of exposing government data to AI systems, rather than building a specific AI tool.
GDS found that The National Archives’ legal data is already close to AI-ready, thanks to high data quality, strong leadership, relevant skills, and mature governance practices. It said that good data alone is not enough; public sector organisations also need the right people, processes, and culture to use data safely, ethically, and responsibly.
The project also identified the evaluation and validation of AI-generated outputs as a significant future opportunity for the government. GDS said public bodies could add value by developing tools and standards to assess whether AI outputs are trustworthy, rather than replicating services already developed by major technology companies.
The next phase will explore how data maturity can reduce the risks of using AI with public sector data. It will also examine technologies such as the Model Context Protocol, an open-source standard for connecting AI applications to external systems, including databases, tools, and documents.
Why does it matter?
The project shows that AI readiness in government depends on more than deploying new tools. Public bodies need high-quality data, strong governance, clear accountability, and the ability to evaluate AI-generated outputs before relying on them in services that affect citizens and businesses. The work also points to a useful role for government: setting standards for trustworthy AI outputs, rather than simply building public-sector versions of commercial AI products.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!