What happened: The UK government announced a new Fundamental AI Research Lab, with up to £40 million over six years plus in-kind access to large-scale compute capacity.

Why it matters: The programme is explicitly aimed at “blue sky” work that tries to fix persistent weaknesses in today’s systems — including hallucinations, short memory, and unpredictable reasoning — rather than simply scaling current approaches.

Wider context: The call is framed as part of a broader UKRI AI strategy backed by £1.6 billion over four years, and it leans on the UK’s research base as a lever for both capability and governance influence.

Background: Applications are open now, with proposals assessed via peer review chaired by Raia Hadsell (Google DeepMind), who is also serving as a UK government AI ambassador.

Government to create new lab to keep UK in the fast lane on AI breakthroughs — GOV.UK

Singularity Soup Take: Funding “fundamental” AI is the right instinct — but governments should be honest that better reasoning and fewer hallucinations aren’t just academic wins; they’re the precondition for deploying AI into high-stakes services without quietly increasing systemic error.

Key Takeaways:

• £40m + Compute: The lab is backed by up to £40 million over six years, plus in-kind access to the AI Research Resource compute capacity valued at tens of millions of pounds.
• Fixing Core Failure Modes: The announcement highlights target problems including hallucinations, unreliable memory, and unpredictable reasoning, aiming for more accurate, transparent, and trustworthy AI systems.
• Open Call, Peer Review: The funding call is open for applications now and will be assessed by a peer review panel chaired by Raia Hadsell, a Vice President of Research at Google DeepMind.