UK's ARIA Puts £22M Into AI-Enabled Formal Cyber Defence
HaydnBelfield · x · 2026-09-24
UK research agency ARIA is funding eight teams with £22m to test whether AI-enabled formal methods can make high-assurance cyber defence practical at unprecedented speed and scale. The effort sits inside its £59m Safeguarded AI programme, which aims to let fleets of AI agents model and verify critical cyber and cyber-physical systems to boost societal resilience.
The programme spans several tracks: TA1.1 covers theory tooling, defining an extendable, interoperable language and formal semantics for world models, probabilistic specifications, neural network controllers and efficiently checkable proof certificates; TA1.2/1.3 cover platform tooling, TA1.4 sociotechnical integration, and TA2 applications. Rolling funding for the Safeguarded AI cybersecurity track (Track 1) remains open.
Related event: UK's ARIA Puts £22M Into AI-Enabled Formal Cyber Defense(3 posts)→
More from Safety
- Gary Marcus weaponizes Jensen Huang's own words: OpenAI's repeated coverups mean we should shut the labs down — GaryMarcus · 2026-09-24
- Jail CEOs or pass bills? The debate over deterring unsafe AI development — trevposts · 2026-09-24
- Anthropic Opened Bio Access and Hired Its Own Referee 11 Days After Asking the World to Slow Down — shaunralston · 2026-09-24
- Anthropic to Host Outside Evaluators On-Site, Other Labs Follow — victor_explore · 2026-09-24
- Wiz Scans US Critical Infrastructure with Gemini-Powered Red Agent — ralucaadapopa · 2026-09-24
- Researchers exploited Cloudflare Containers flaw to read other tenants' residual disk data — matthew_d_green · 2026-09-24