DICE Program

DICE Program is sponsored by Defense Advanced Research Projects Agency (DARPA). The DARPA DICE program seeks to develop AI collectives that are powerful, predictable, aligned, and resilient by harnessing principles of self-organizing systems and rigorous inference-time control.

Department of War organization. Decentralized Artificial Intelligence through Controlled Emergence Future conflicts will unfold at machine speed in highly dynamic and contested environments. This will require autonomous, multi-agent artificial intelligence (AI) systems to create an asymmetric battlespace advantage and reduce risk to warfighters.

The Decentralized Artificial Intelligence through Controlled Emergence (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 our 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. DICE aims to develop a decentralized AI architecture suitable for rapidly evolving, unpredictable, and contested environments.

With this architecture, AI agents can dynamically form teams using peer-to-peer coordination to execute complex missions. This coordination will be robust to failure or compromise of individual agents, as well as to “rogue” AI agents that might develop misaligned instrumental goals.

The local inference control on each AI agent will ensure role coherence of individual agents and constrain the emergent behavior of the collective to maintain alignment with commander’s intent over the long term, even across multiple inference steps.

DICE will seek to expand and leverage the theory of self-organizing systems and distributed consensus algorithms, together with recent breakthroughs in controlling the internal reasoning and inference of AI foundation models at inference-time. DICE aims to advance AI beyond individual agent capabilities by fostering system-level capability that emerges from the interactions between agents.

This controlled emergence ensures DICE AI agents remain on mission, maintain doctrine, suppress misbehaviors, and remain resilient to agent loss or compromise. The program’s scope does not include the development and deployment of autonomous systems in the real world.

The program will use simulation environments to demonstrate DICE architectures in Department of War-relevant use-cases, targeting measurable gains in scalability, adaptability, and resilience against both benign failures and adversarial attacks. Information Processing Techniques Office Proposers Day presentation