The UK government has brought a small team of specialist technologists into Whitehall to build open-source artificial intelligence (AI) tools for transport, public safety and defence, with funding provided by Meta.
The recruits will spend the next year embedded across government, working directly with departments rather than external vendors. Their brief is to design systems the state can run itself, without handing sensitive data to closed, commercial platforms.
This work sits under the $1 million “Open-Source AI Fellowship”, announced in July 2025 and run through the Alan Turing Institute.
The fellowship places experienced engineers and researchers inside public offices for 12 months, with a mandate to deliver practical tools, not policy papers.
The team brings together skills from academia and applied research, including data science, computer vision, robotics-led imaging and the design of safety-critical systems.
Some members come from the Alan Turing Institute, others from UK universities. Their focus is not experimentation for its own sake, but solving everyday government problems.
Transport is one of the early targets. Officials want systems that can flag road damage before it becomes dangerous, predict maintenance needs and ease pressure on overstretched networks.
Public safety work is expected to centre on faster analysis of incidents and better coordination during emergencies. Defence and national security teams are exploring tools to support decision-making, translation and threat detection.
What makes the programme different is ownership. Tools developed through the fellowship will belong to the government. Departments will be able to adapt them internally and keep sensitive information in-house.
The software will be open-source, a development intended to cut expenses, improve transparency and reduce long-term dependence on proprietary systems.
Meta, which is funding the initiative, has said the work will draw on its Llama models, which can handle text, images, audio and video. The government, however, retains control over how the technology is deployed and modified.
This aligns with Prime Minister Keir Starmer’s l economic agenda, which treats artificial intelligence as a lever for productivity across public services while keeping a tight grip on security risks. It also places the UK on a distinct path internationally.
Europe is tightening regulations through the EU AI Act, while the United States has set up a dedicated AI Safety Institute. Britain is placing itself somewhere in between, betting on open systems that are powerful, auditable and firmly under state control.
