The Alan Turing Institute has proposed building a British AI model from scratch and making it freely available to the public.
The idea appears in advice published by the government’s Council for Science and Technology (CST), which wants ministers to consider the model as open, UK-governed national infrastructure. It would be available to researchers, start-ups, small businesses, public services and the open-source community, rather than being controlled by one company or department.
A broader plan for UK AI
The CST has recommended allocating some national AI computing capacity to test different types of models, including open-weight systems and smaller domestic models. The trials would explore practical uses in areas including health, education and national security.
The council also called for a UK AI Sovereignty Plan, mapping the country’s computing power, data, skills, models and governance. It wants ministers to decide which capabilities Britain should own, share or secure access to, while identifying important datasets and testing access arrangements in areas such as health and creative data.
Other recommendations include appointing a single senior owner for AI sovereignty, creating an AI skills and secondment route alongside the TechFirst programme, and continuing support for low-power computing research.
Why it matters for Britain
A publicly available UK model could give researchers, smaller firms and public services an alternative to relying entirely on overseas systems. The proposed pilots could also help develop domestic skills and test where British-built AI is most useful.
