ARIA funds £22m program; Keystone to build formally verified secure LLM serving infrastructure
mboehme_ · x · 2026-09-24
UK's ARIA is funding eight teams with £22m to test whether AI-enabled formal methods can make high-assurance cyber defense practical at unprecedented speed and scale, in response to faster and farther-reaching AI-enabled cyber attacks.
Among them, afdicl (Imperial College) will lead Keystone: using formal methods to build highly secure LLM serving infrastructure, leveraging AI itself to scale. The author calls this a golden age for formal methods.
Related event: UK's ARIA Puts £22M Into AI-Enabled Formal Cyber Defense(3 posts)→
More from Safety
- X Exec's AI Refund-Scheme Prompt Branded Fraud — Even Grok Agrees — PTrubey · 2026-09-25
- Attendee at King Charles' AI convening: builders failed to commit to adequate principles — BlackHC · 2026-09-25
- Google, OpenAI and Anthropic reportedly forming their own frontier-AI safety authority — 141_1337 · 2026-09-25
- Meta's Muse AI agent tricked into sharing its entire filesystem with minimal prompting — The Verge AI · 2026-09-25
- Hugging Face CEO: AI risk comes from secret frontier labs, open-source is the fix — ivan_bezdomny · 2026-09-25
- Dev Hooks an LLM to a Robotic Car, Strips Safety Rules, Cites His LLC — ostrisai · 2026-09-25