Google DeepMind · 2026-06-11 · major
DeepMind Opens $10M Multi-Agent AI Safety Research Grant
Google DeepMind, Schmidt Sciences, Cooperative AI Foundation, ARIA, and Google.org open a $10M funding call for research on agent-to-agent safety. Four priority areas, applications due August 8.
A $10M research call for tools to keep millions of AI agents safe when they start talking to each other.
Key specs
| Funding pool | $10M |
|---|---|
| Partner orgs | 5 |
| Application deadline | 2026-08-08 |
What is it?
Google DeepMind and four partners — Schmidt Sciences, Cooperative AI Foundation, ARIA, and Google.org — are funding outside research on multi-agent AI safety. Most current safety work studies one model at a time; this grant targets the new problem of many agents from different vendors interacting in shared environments.
How does it work?
The program funds work across four areas: sandboxes and testbeds for multi-agent evaluation, the science of agent networks and emergent behavior, secure agent infrastructure protocols, and oversight and control mechanisms for deployed agent populations. Applications go through Schmidt Sciences' portal; deadline is August 8, 2026.
Why does it matter?
Coding agents, browser agents, and customer-support agents are already in production. Once they start trading messages with each other across companies, single-model safety evaluations stop covering the risk. A $10M pool with top-lab branding helps pull academics into a young area that has been mostly empty.
Who is it for?
academic AI safety researchers, multi-agent systems labs
Try it
https://schmidtsciences.smapply.io/prog/scaling_ai_safety_for_a_multi_agent_world