Un new report from The Alan Turing Institute from the United Kingdom, reveals that up to 41% of the time spent on tasks in the public sector could be optimized or improved through the use of generative AI. This proportion varies by field: 49% in Education, 33% in Health, and is higher among staff doing administrative tasks (47%) than among staff doing citizen service (38%).
The objective is to identify real opportunities for strategic and efficient adoption of this technology, which is already transforming public management. This work representa an innovative contribution to the map of AI capabilities in the public sector and opens the door to reflecting on how we can apply these learnings in our context.
Optimizing is not replacing or eliminating
The report does not suggest that generative AI should eliminate functions or replace people. On the contrary, it highlights the need for robust adoption strategies to ensure that the technology is integrated in a responsible and beneficial way. When it talks about optimizing or supporting the tasks of public employees, it refers to streamlining repetitive or administrative processes, improving quality or operational efficiency, or freeing up time for public professionals to devote to more value-added tasks.
Exclusive
The study is based on the Public Sector Time Use Survey. The methodology consists of analyzing the specific tasks performed by public employees. These tasks have been compared with the current capabilities of generative AI to determine which ones could be automated or assisted. The average time spent on each activity has also been calculated to estimate the overall optimization potential.
More information
Hashem, Y., Bright, J., Chakraborty, S., Onslow, K., Francis, J., Poletav, A., Esnaashari, S. (2025). Mapping the Potential: Generative AI and Public Sector Work. Using time use data to identify opportunities for AI adoption in Great Britain's public sectorThe Alan Turing Institute.
Illustration: “Policymakers” by Jonny Lighthands, Turing Commons, Licensed under CC BY-SA 4.0
