By Vikas Kanungo, AI & Digital Transformnation Leader, July 16, 2025
In a landmark move aimed at strengthening AI capacity within the public sector, the United Kingdom has launched a $1 million Open-Source Generative AI Fellowship. The initiative, backed by the UK government in collaboration with Meta and the Alan Turing Institute, will begin in January 2026 and invites top AI engineers to embed directly within government departments. These engineers will co-develop open-source generative AI tools using Meta’s Llama 3.5, with a strong emphasis on real-world governance challenges and sovereign technology development. The Fellowship is one of the first of its kind globally to combine generative AI development with embedded public service delivery at scale.
The program’s objective is to empower public institutions with in-house capabilities to design, train, and deploy large language models tailored to core public service needs. Fellows will work on transformative applications such as real-time translation tools for national security, AI-powered assistants to streamline housing permits, and resilient emergency response systems that function even in low-connectivity environments. These solutions build on the UK’s earlier pilots like “Caddy,” a customer service chatbot that improved response times by over 50%, and “Humphrey,” an administrative copilot already deployed across several departments. Importantly, all tools developed under the Fellowship will be open-sourced, ensuring transparency, adaptability, and reusability across the broader AI in governance ecosystem.
AI practitioners participating in this initiative will gain first-hand exposure to the complexities of deploying generative AI in high-stakes environments such as planning, emergency services, and citizen engagement. The Fellowship offers a rare opportunity to apply frontier AI skills in the public domain, contributing not just code but also frameworks for responsible and scalable use. With direct access to secure data environments and policy teams, fellows will help shape how AI integrates into the fabric of modern governance. Applications are now open via the UK Government’s official portal, with the Alan Turing Institute leading the candidate selection and onboarding process.
For governments around the world, especially countries like India, the UK’s Fellowship provides a timely model for India AI mission. The program exemplifies how open-source LLMs and local AI engineering talent can drive generative AI adoption in a way that is sovereign, transparent, and sustainable. Instead of relying exclusively on external vendors, the UK is creating a home grown AI workforce that understands both the technical and policy dimensions of public sector AI integration. This embedded talent model is particularly relevant for India’s own IndiaAI Mission and its support for domestic LLMs like those from Sarvam AI and BharatGPT. The UK Fellowship shows that building internal capabilities, backed by strong public-private-academic partnerships, is not only feasible but also critical to long-term digital sovereignty.
This initiative also delivers three core lessons for the global AI and governance community. First, it underscores the importance of treating generative AI tools as digital public goods, built and governed in the public interest. Second, it validates open-source LLM development as a practical pathway toward trustworthy, auditable AI systems in complex policy environments. And third, it highlights how structured capacity-building—through fellowships, knowledge hubs, and prompt libraries—can accelerate institutional readiness at scale. The UK’s Open-Source AI Fellowship is not just a national program; it is a blueprint for how governments worldwide can responsibly harness the power of generative AI to serve their citizens better.
