Exploratory Models of Human-AI Teams (EMHAT)

Background
The Defense Advanced Research Projects Agency (DARPA) is issuing an Artificial Intelligence Exploration (AIE) Opportunity, inviting submissions of innovative basic or applied research concepts in the technical domain of Exploratory Models of Human-AI Teams. The solicitation emphasizes the increasing interdependency between humans and AI, and the need for comprehensive modeling and simulation toolkits capable of addressing complex, often emergent behaviors of groups composed of humans and generative AI systems.

Work Details
The Exploratory Models of Human-AI Teams (EMHAT) will leverage generative AI to create a human-AI modeling and simulation framework capable of providing sufficient naturalistic data to enable quantitative modeling and assessment of the task execution effectiveness of human-machine teams in realistic settings. The technology will utilize expert feedback, AI-assembled knowledge bases, and generative AI to produce computational agents representing a diverse set of human teammate simulacra, analogous to digital twins. Proposals should address two independent and sequential project phases: a Phase 1 Feasibility Study (base) and a Phase 2 Proof of Concept (option), with periods of performance not exceeding 18 months. The Phase 1 award value is limited to $600,000, and the Phase 2 award value is limited to $400,000.

Period of Performance
The period of performance for the Phase 1 base effort is 9 months, followed by a 9-month period for the Phase 2 option effort. The combined Phase 1 base and Phase 2 option efforts should not exceed 18 months.