To see real results from the government’s AI strategy, strong leadership and skill development are crucial, along with the recognition that outdated IT systems need a refresh. This message came through clearly during a Public Accounts Committee session on January 30, led by Sarah Munby, the permanent secretary for the Department for Science, Innovation and Technology.
Munby pointed out that self-criticism is tough for organizations, but it’s necessary to improve digital leadership across the board. She referenced Deloitte’s report, highlighting that many leaders in the public sector lack the technical know-how to drive tech initiatives forward, especially when it comes to opportunities like AI, which could be vital to their missions.
Catherine Little, the chief operating officer for the civil service, echoed the sentiment about the skills gap. She mentioned that both public and private sectors are struggling to find the right experts to implement these technologies effectively. Everyone is competing for the same talent, which presents a big challenge for central government.
David Knott, chief technology officer of the Central Digital and Data Office, discussed the varying technical hurdles different departments face. He indicated that there have been consultations with departments to pinpoint specific obstacles, revealing that skills shortages remain an issue and that data needed for training AI models is often stuck in outdated systems.
Knott admitted that progress on updating these legacy systems hasn’t been even across the board. He also touched on the maturity of the market for AI products, noting that many of the solutions being developed commercially need careful evaluation before the government can adopt them. There’s a balancing act between central procurement and allowing smaller companies to offer localized solutions.
Finally, Knott pointed to change management as a significant challenge related to AI. With AI tools becoming part of everyday workflows, the shift in working habits can be dramatic, similar to moving from command line interfaces to graphic user interfaces in computing.