The BBC reported on 23 October that the UK will set up a national AI Biosecurity Lab to strengthen its defences against new biological threats. Wednesday's Budget will commit £100m to it. OpenAI and Anthropic will supply £200m in AI model credits, which can be spent on using their models. The Budget also adds £120m for the AI Security Institute, which tests risks in advanced AI.
The BBC said the lab's role and remit are not yet known. It does not say when the lab opens, who may spend the credits, or on what. The government also named two cyber plans. One is £60m for a National Cyber Shield using agentic AI, meaning AI that carries out tasks on its own. The other is £70m for a centre that uses AI tools to find weak spots in critical systems. The BBC noted these came after the UK took some sites offline over php2shell, a flaw in PHP. It reports no successful exploits so far.
The credits are worth twice the government's £100m, and they pay for using models rather than providing cash. Experts told the BBC the PHP case shows how fast AI can surface long-dormant bugs. Buyers may ask what the credits cover and who sets the terms.