The government's ambition to embed artificial intelligence across the public sector is at risk due to outdated technology, poor data quality, and a lack of skilled staff, according to a report by the cross-party Public Accounts Committee (PAC). The report highlights that over 20 government IT systems identified as 'legacy'—meaning out of date and unsupported—have yet to receive funding for replacement.
Government research cited by the PAC found that almost a third of central government IT systems met the legacy definition in 2024. Prime Minister Keir Starmer has repeatedly stressed the desire to boost economic growth through AI adoption, including in the public sector, and has called for AI to replace government officials where possible, alongside recruiting 2,000 new tech apprentices.
The PAC report also warns of 'persistent digital skills shortages' in the public sector, partly due to civil service pay rates that are uncompetitive with the private sector. The Department for Science, Innovation and Technology (DSIT), which leads AI roll-outs, has acknowledged similar concerns in earlier reports, but the PAC makes new recommendations, including a six-month deadline for DSIT to outline funding for replacing the highest-risk legacy systems and to assess the costs of inaction.
Further recommendations include improving public confidence in AI transparency, as only 33 official records on algorithm-assisted decisions had been published by January 2025. The report also calls for coherent systems to learn from various AI pilots across government. PAC chair Sir Geoffrey Clifton-Brown said the government wants to 'mainline AI into the veins of the nation' but questioned whether the public sector is ready, urging that digital professionals be included in senior management boards.
A government spokesperson responded that the findings reflect known issues, and a bold plan is already underway to overhaul tech and AI use, including doubling tech experts in Whitehall and replacing legacy IT systems more quickly.
