…OpenAI and Anthropic will provide £200m in AI model credits as Britain expands investment in artificial intelligence safety, cybersecurity and protection against biological threats.
The United Kingdom government is set to invest £100m in a new Artificial Intelligence Biosecurity Lab, described as a “world first”, as part of a broader strategy to strengthen national defences against emerging biological, cyber and AI-enabled threats.
The funding commitment, to be announced in Wednesday’s Budget by Chancellor John Healey, will support the establishment of the National AI Biosecurity Lab (AIBL), bringing together leading scientific expertise and advanced AI technology to identify and counter potential biological risks.
AI developers OpenAI and Anthropic will support the initiative by providing £200m worth of AI model credits, which can be used to access their models.
The government has yet to disclose the lab’s full remit, operational structure or timetable, but says the institution will help Britain respond to evolving security challenges as increasingly capable AI systems become more widely available.
AI Minister Kanishka Narayan described the proposed facility as a major new global institution that would combine world-leading scientific expertise to protect the UK’s citizens and national interests.
The investment reflects growing government concern that advances in artificial intelligence could create new security vulnerabilities even as the technology offers opportunities to improve productivity, scientific research and public services.
Healey said the government wanted an approach to AI that protected national interests and ensured workers were not left behind, while spreading the benefits of AI-enabled economic growth across the country.
The biosecurity lab forms part of a wider package of planned spending on artificial intelligence safety and cybersecurity.
The Budget will include an additional £120m for the UK’s AI Security Institute (AISI), which was established in 2023 under the previous Conservative government.
The institute evaluates the capabilities and risks of advanced AI models and has played a role in developing collaborative approaches to testing their safety and security.
The additional funding is intended to sustain and expand its work as governments and technology companies face growing calls for stronger technical standards, safeguards and oversight of increasingly powerful AI systems.
The government will also commit £60m to a proposed National Cyber Shield, which is expected to use agentic AI to strengthen the country’s cyber defences.
A further £70m will go towards a National Exposure and Real-Time Vulnerability (NERV) Centre. The centre is intended to use AI tools to identify vulnerabilities across critical infrastructure and the wider economy.
Together, the measures signal a growing emphasis on using AI not only to drive innovation but also to detect and respond to threats that may be difficult to identify through conventional security systems.
The investment comes amid concerns about the potential use of AI to discover and exploit weaknesses in widely used software.
The UK government recently joined organisations around the world in taking some websites offline as a precaution after researchers identified serious vulnerabilities in PHP, a programming language widely used in web development.
Researchers warned that the flaws could expose affected platforms to hacking. The vulnerability, referred to as “php2shell”, had not been linked to any reported successful exploitation at the time of the report.
However, the incident highlighted concerns that AI tools could accelerate the discovery of long-standing software weaknesses, potentially reducing the time available for organisations to identify and fix vulnerabilities before attackers exploit them.
The proposed National Cyber Shield and NERV Centre are intended to improve the ability to detect such weaknesses and respond more quickly.
The challenge for governments and businesses will be to ensure that defensive AI systems can keep pace with the same technology being used by malicious actors.
