Meta and the UK Government have announced a joint initiative to advance artificial intelligence in the public sector, launching a £1 million Open Source AI Fellowship in partnership with the Alan Turing Institute. Set to begin in January 2026, the 12-month programme will embed AI specialists within government departments to develop tools designed to improve public services and support national security.
Described as a step towards becoming an “AI maker not an AI taker,” the programme will use open-source models, including Meta’s Llama series, ensuring government ownership of all outputs. This approach protects sensitive data, avoids vendor lock-in, and allows AI tools to be tailored to specific public sector needs. “This is the best of AI in action—open, practical, and built for public good,” said Peter Kyle, Technology Secretary. He emphasised the government’s focus on delivery and its ambition to build tangible systems that improve services and operations.
Dr Jean Innes, CEO of the Alan Turing Institute, called the fellowship a novel way to match AI expertise with real-world public challenges. Fellows will work on projects such as real-time language translation for national security and using planning data to speed up home building approvals. The programme aims to unlock up to £45 billion in public sector productivity.
The fellowship builds on existing tools like ‘Humphrey’, an AI assistant helping civil servants summarise documents, take meeting notes and compile consultations. Another example is ‘Caddy’, developed with Citizens Advice, which has halved call response times and doubled advisor confidence. Caddy has since been open-sourced for wider use in services like debt advice and legal support.
Participants will join the Department for Science, Innovation and Technology’s Incubator for AI—the team behind Humphrey—to help turn experimental ideas into operational solutions. The fellowship complements the government’s AI Knowledge Hub, which shares case studies, tools and a Prompt Library to help teams scale responsible AI use.
Joel Kaplan, Meta’s Chief Global Affairs Officer, said the programme reflects a shared commitment to open, secure and sovereign AI development. By bringing together private innovation and public service goals, the fellowship offers a potential model for how open-source AI can address complex challenges while upholding public trust.
Applications will open soon, inviting AI experts keen to contribute to the next phase of public sector transformation. Through this initiative, the UK is setting a strong example in the responsible use of AI to modernise government and improve the lives of its citizens.
Created by Amplify: AI-augmented, human-curated content.
