UK sets out additional £120 million for AI Security Institute
The UK government has outlined an additional £120 million for the AI Security Institute to expand model testing, risk research and technical-standards work.
The UK government has set out an additional £120 million for the AI Security Institute, proposing to use the investment for testing advanced AI systems, researching their capabilities and risks, and developing technical standards.
The Treasury published the funding line on October 24 as part of a broader AI investment package tied to the Chancellor’s Budget plans. According to the government, the allocation would give the institute more resources to test new AI systems and build technical understanding of how their risks can be mitigated.
The AI Security Institute is a research organisation within the Department for Science, Innovation and Technology. It tests leading AI systems, briefs policymakers on emerging capabilities and risks, and researches possible safeguards. The institute’s research agenda says it also turns findings into safety standards and protocols and works with frontier-model developers as an independent technical partner.
The £120 million is specifically for AISI, separate from the package’s £165 million National AI Biosecurity Initiative. That initiative comprises £100 million for an AI Biosecurity Lab and £65 million for biosurveillance and medical countermeasures for priority pathogens.
The government listed two distinct cyber-defence programs as well: £60 million for a National Cyber Shield, including a £25 million challenge fund, and £70 million for a National Exposure and Real-Time Vulnerability Centre intended to use AI tools to identify cyber vulnerabilities. Neither amount is part of AISI’s allocation.
The government’s sovereign-AI plans contain two further £60 million allocations. One would add capital to the Sovereign AI Fund for direct investment in UK AI companies. The other would support a research-and-development procurement scheme covering health and care, AI security, assurance and safe adoption. Both are separate from the AISI funding.
The new amount follows £240 million that the government said it committed to AISI at the 2025 Spending Review for frontier-model testing, foundational safety and societal resilience. The October release does not explain how the additional £120 million changes that earlier funding total or period.
The Treasury page uses mixed timing language: it says the package would be announced the following week, but also says the Chancellor had announced the investments. It does not specify the AISI money’s spending period, disbursement schedule or division among testing, research, staffing, infrastructure and standards work.
More news

OpenAI to require approval for model-access intermediaries

Anthropic pitches lawmakers on draft frontier AI safety bill

OpenAI says research model spent funds from an unauthorized test wallet
