The UK government is putting £100 million into a new AI Biosecurity Lab, and it's recruiting OpenAI and Anthropic to help run it. The Cabinet Office announced the project on Saturday, calling it the world's first dedicated organisation built to use artificial intelligence against biological threats.
The lab will start with two jobs: catching new biological threats before they spread, and speeding up the hunt for vaccines and treatments. Neither OpenAI nor Anthropic is writing a cheque, though. OpenAI's contribution, worth up to £100 million, comes as compute and model credits plus access through its Rosalind Biodefense programme. Anthropic is chipping in a matching amount in model credits, model access and technical support. Together the two firms could add up to £200 million on top of the government's £100 million, all of it in kind rather than cash.
Google DeepMind is also joining the research effort and may open up access to its Gemini and science-focused models. A UK company, Basecamp Research, is contributing its own AI biology models to the mix.
Who runs it, and what's still missing
The lab will operate as an independent non-profit, structured the way the AI Security Institute already is. What the Cabinet Office hasn't said yet is almost as notable as what it has: no director has been named, no site has been picked, and there's no public timeline for how long the £100 million government stake needs to last.
A handful of public bodies are lined up to take part, including the UK Health Security Agency and the Defence Science and Technology Laboratory. The medicines regulator, the MHRA, will work alongside the lab specifically on how to regulate drugs that AI systems help design, a question regulators worldwide are still working out.
The biosecurity lab is the headline item in a wider £600 million-plus package for AI security that the government laid out ahead of Wednesday's Budget. That package also includes £120 million for the existing AI Security Institute, according to City AM. Chancellor John Healey announced the funding, and AI minister Kanishka Narayan described the new lab as a major global institution, the Evening Standard reported.
The arrangement fits a pattern that's becoming familiar in AI policy circles: governments announcing nine-figure sums that turn out to be compute and model access rather than money changing hands. Nvidia did something similar earlier this month, pledging $250 million worth of compute to UK AI safety research rather than cash.
For OpenAI and Anthropic, the deal buys goodwill and a formal seat at the table on biosecurity policy, an area where both companies have faced questions about whether their models could help someone design a dangerous pathogen as easily as they could help stop one. For the UK, it's a way to stand up a flagship AI safety institution without drawing much directly from the Treasury. What's missing so far is the basic infrastructure, a building, a leader, a start date, that would turn the announcement into an operating lab.
TBFiled by The Tech Business Desk
Big tech, startups, earnings, regulation and policy — the decisions that shape the products.
More from this desk →
Be the first to comment
Join the argument. No password, just your email or a passkey.