The UK government has allocated £60m to establish two new AI research labs at the University of Oxford and University College London, aiming to make artificial intelligence cheaper, more accessible, and better aligned with human needs. The investment, announced on 23 June 2026, doubles the previously planned number of labs and forms part of UK Research and Innovation's £1.6bn AI strategy.
Labs focus on open-source AI and efficient learning
Sofair, the Science of Fundamental AI Research Lab, will be led by Professor David Barber at UCL, with contributions from Cambridge, Oxford, and Edinburgh. The lab aims to develop open-source AI technologies that run on widely available hardware. Researchers from computer science, mathematics, statistics, and neuroscience will explore new design approaches to lower costs and reduce dependency on a small number of model providers.
The British Open-ended Learning and Discover Lab (Bold), led by Professor Jakob Foerster at the University of Oxford, will rethink how AI learns from its environment. Supported by UCL and Imperial College London, Bold will develop systems that learn more efficiently, adapt to new situations, and navigate physical spaces. Its focus is on practical, human-centred AI for workplaces, infrastructure, and public services.
For scientists looking to build expertise in these areas, AI for Science & Research training provides a pathway to develop the skills needed to contribute to such developments.
Funding and institutional support
The funding will be distributed over six years through UKRI's Engineering and Physical Sciences Research Council (EPSRC). Both labs will receive £2m to hire at least 10 doctoral students each. They will work closely with the Alan Turing Institute and UKRI's nine AI research hubs, and will have access to large-scale computing power. The announcement adds £20m to the original plan, doubling the number of labs from two to four.
Voices from the labs
Professor Foerster said: "The UK cannot win the global AI race simply by trying to outspend the largest technology companies on data and compute. Bold is about a different route: discovering fundamentally new ways to build AI that are more efficient, more open and better aligned with human needs."
Professor Barber said: "While current AI systems are impressive, many still suffer from basic issues such as inaccurate responses to questions. These systems often use similar underlying architectures, so Sofair will bring together the broader sciences and fresh ideas to create a new generation of open-source models. This will reduce dependency on the small number of model providers, boosting UK sovereignty and its position as a global player in AI."
AI minister Kanishka Narayan said: "By building this capability here at home, backed by our world leading universities, we're strengthening our own expertise, reducing reliance on others and securing Britain's place at the forefront of this technology - fittingly announced on what would have been Alan Turing's 114th birthday."
Why this matters for science and research professionals
The labs will create a pipeline of open-source tools and new learning methods that directly impact how AI can be applied in research settings. The doctoral studentships and focus on fundamental breakthroughs mean more researchers will enter the field with skills in building affordable, adaptable AI. For laboratories, clinical research units, and data-intensive scientific teams, the shift towards models that can learn from less data and run on common hardware could reduce costs and increase reproducibility. The emphasis on human-centred, interpretable systems also promises AI that supports scientific discovery rather than remaining a black box.
Your membership also unlocks:
