Funding, Scaling AI Safety for a Multi-Agent World

This funding call aims to support research that ensures safety in a world with millions of interacting AI agents. Key areas include:

• Creating sandboxes and testbeds to study frontier-model agents with realistic tools, memory, and communication.
• Analyzing how individual agent properties contribute to system-level safety outcomes and emergent vulnerabilities.
• Strengthening agent infrastructure for identity, authentication, reputation, and commitment.

This research will help ensure that the deployment of millions of interacting AI agents is safe and beneficial, preventing catastrophic outcomes from emergent vulnerabilities.

Researchers and teams with expertise in:

• AI safety, multi-agent systems, or related fields.
• Building testbeds or simulation environments for AI agents.
• Infrastructure development for identity, authentication, or reputation systems.

Funding for research projects. Specific compensation details are not provided in the posting.

Schmidt Sciences is an organization that supports scientific research and innovation, with a focus on AI safety and other high-impact areas.