$1 million AI fellowship to innovate public services
AI

AI experts can now apply for a 12-month fellowship, backed by $1 million, to build open-source AI tools that improve public services, boost productivity, and support national security.

 

The “Open-Source AI Fellowship” has been funded by a grant from Meta to the Alan Turing Institute, with fellows set to join the Department for Science, Innovation and Technology’s (DSIT) Incubator for AI, the team behind “Humphrey,” which are a collection of AI tools to help civil servants. Fellows will have the opportunity to expand Humphrey, by coming up with innovative solutions to further relieve civil servants of their administrative burdens.

 

Additionally, fellows could join government to build AI tools for high-security use cases across the public sector such as language translation in a national security context, and making use of construction planning data to speed up the approvals process and get more homes built.

Beginning in January 2026, the fellowship will last twelve months, and applicants are able to register their interest ahead of applications going live next week.

 

The fellows will be primarily deployed on open-source AI models, which could reduce costs to the taxpayer when using AI widely, and help unlock up tot £45 billion in productivity gains across the public sector.

Technology secretary Peter Kyle said: “This Fellowship is the best of AI in action—open, practical, and built for pubic good. It’s about delivery, not just ideas—creating real tools that help government work better for people.

 

“We’ve already seen the potential. Caddy—developed with Citizens Advice and now helping Cabinet Office teams—shows how open AI tools can boost productivity, improve decision-making, and support frontline staff.

 

“The Fellowship will help scale that kind of impact across government, and develop sovereign capabilities where the UK must lead, like national security and critical infrastructure.”