Decentralized Artificial Intelligence through Controlled Emergence (DICE)
DARPA seeks research proposals to develop theory and algorithms for decentralized coordination and local inference control for scalable, resilient multi-agent AI systems under human control.
⚑ DARPA solicitation; federal contract mechanism. · Applicant restrictions are not further specified in the notice beyond federal contract eligibility.
This is a DARPA AI/computing program focused on decentralized agent algorithms and controlled emergence, which is far outside IPPRA’s core social-science research lines. The team’s defense-adjacent work is about public risk perception and communication, not AI systems design or distributed autonomy.
JUDGED AGAINST THE ROSTER'S PUBLICATION DOSSIER · GPT-5.4-MINI · 2026-07-07
Unit fits — one characterization, each unit's own rules
| Physical Sciences & Engineering (demo) | 90 strong | technical depth: central; funds basic research |
| IPPRA | 58 good | portfolio topics: national_security_defense, emergency_disaster_resilience; social/behavioral work is minor; funds basic research |
| Tom Love Innovation Hub | 30 weak | prototyping/demonstration stage; deep-tech content |
Description
The DICE program seeks to develop the theory and algorithms for decentralized coordination and local inference control to enable a scalable, adaptive, and resilient collective of heterogeneous AI agents that can autonomously execute sustained long-time-horizon missions in contested environments while remaining under human control. In contrast to small-scale, rigid, and fragile centralized orchestration or the high-risk unpredictable nature of ad hoc compositions of AI agents, DICE aims to harness the scalability and adaptability of self-organizing systems while minimizing risks and ensuring that the collective behavior remains predictable and aligned with intended outcomes. This approach mirrors the principles of decentralized self-organization that underpin the internet's own scalability and resilience, where robust global behavior emerges from simple, local rules.
Eligibility
Applicant restrictions (federal contract). Set-aside: . Notice type: Solicitation. Organization: DEPT OF DEFENSE / DEFENSE ADVANCED RESEARCH PROJECTS AGENCY (DARPA) / DEF ADVANCED RESEARCH PROJECTS AGCY.
Apply
Proposal brief SEE AN EXAMPLE →
A one-page internal memo: fit assessment, submission requirements, document scaffold, and next steps dated back from the deadline — tailored to your project idea if you add one.
Proposal shell · Department of Defense (BAA-style) conventions SEE A DOD EXAMPLE →
Funder-faithful document skeletons — Department of Defense (BAA-style)'s document set with section headings, page limits, reviewer guidance, and writing prompts; add a project idea to get [DRAFT] starter bullets. Download as .md for Word or Overleaf.