UK ARIA launches £50m Scaling Trust programme funding secure AI agent coordination
HaydnBelfield · x · 2026-10-08
ARIA (UK) unveiled the first cohort of its near-£50m Scaling Trust programme, funding teams building infrastructure for AI agents to coordinate securely on our behalf across digital and physical worlds.
- Track 2 (open-source cyber-physical agent stack): secure agentic interactions, TEE sandboxes, watermarks embedded in robotic movements, autonomous manufacturing data, plus an integration partner building an end-to-end app.
- Track 3 (fundamental theory): two teams exploring cryptography and complexity theory approaches — confidentiality built into AI itself, and 'self-proving' models that emit checkable proofs alongside answers.
Everything is open-source by default, with tools competing in an Arena under adversarial pressure. Rolling quarterly grants will add new teams.
Related event: UK ARIA Commits Nearly £50M to AI Agent Safety Research(2 posts)→
More from Safety
- Security researcher: the best AI bugs live at the safety-security intersection — wunderwuzzi23 · 2026-10-08
- Six years on, wunderwuzzi's ML attack series on model backdoors and Assume Breach still holds up — wunderwuzzi23 · 2026-10-08
- EMNLP 2026 paper RECAP trains reasoning models to recover from unsafe trajectories — pinyuchenTW · 2026-10-08
- SaTML 2027 to host AdvML Frontiers workshop on human-centered trustworthy machine learning — pinyuchenTW · 2026-10-08
- Nature warns discreet AI wearables risk normalizing non-consensual recording; Subtle founder responds — tianshi_li · 2026-10-08
- Requesting Anthropic's CVP auto-suspends your account, appeal restores it — dyn___ · 2026-10-08