The U.S. Defense Advanced Research Projects Agency (DARPA) has launched a new program to solicit paradigm-shifting research ideas to revolutionize how scientists predict collective human behavior.
The program, known as Methodological Advancements for Generalizable Insights into Complex Systems (MAGICS), aims to address the problem that despite the rise of big data and machine learning, we’re still surprisingly bad at forecasting how large, dynamic human systems respond to change.
“For the past decade or more, there has been an assumption and hope that the explosion of digital data streams (e.g., social media, purchase patterns, traffic dynamics, etc.) combined with powerful machine learning tools would usher in a new era of research in complex, dynamic, evolving systems,” a DARPA solicitation notice writes. “[However] Despite many attempts, results have failed to meet expectations.”
The MAGICS opportunity, announced through DARPA’s Defense Sciences Office, invites individual researchers to propose innovative concepts that could form the foundation for a new science of social prediction.
As DARPA notes, today’s best statistical tools often falter when applied to real-world, evolving systems—whether it’s understanding how economies adapt to disruption, how populations shift under demographic pressure, or how societies react to technological upheaval.
At the heart of the MAGICS effort is answering the question: Can we develop new ways to model collective human behavior that outperform current statistical approaches and capture the dynamics of complex, evolving systems?
The Pentagon brain trust is looking for fresh frameworks beyond what’s possible with today’s machine learning models, namely systems that can handle the messy, recursive, and often unpredictable nature of human systems.
The stakes are high for national security. From forecasting the spread of misinformation to anticipating societal responses to crises, the ability to model human behavior accurately could offer profound advantages.
Yet DARPA acknowledges that researchers must overcome foundational challenges that large datasets and artificial intelligence have failed to address before these benefits can be realized.
Among these challenges are what DARPA calls “unstable mappings” between the hidden psychological factors driving behavior and the observable data points researchers try to analyze.
There’s also the problem of generalization: models that work well in one context often fall apart when applied elsewhere. In many cases, existing tools struggle to account for how systems change in response to being observed—a phenomenon familiar to anyone who has seen how public scrutiny can alter behavior.
DARPA Emphasizes the need for new thinking to address the research gaps. This includes deriving meaning from sociotechnical data sets, developing new techniques, gaining theoretical insights, and understanding the fundamental limits of inference from available data.
In solicitation documents, DARPA makes it clear it is asking proposers to think big. The agency isn’t interested in incremental tweaks to existing techniques. Instead, it seeks entirely new methodologies, metrics, and theoretical foundations that can handle social systems’ inherent uncertainty and reflexivity.
To that end, MAGICS proposals must combine theoretical, computational, and empirical elements. They should draw from various disciplines—including psychometrics, behavioral science, data science, and machine learning—while avoiding narrow approaches that address only a single facet of the problem.
For example, improving a survey instrument or refining a single algorithm will not be enough unless the work is tied to a larger, generalizable framework that connects to real-world phenomena.
DARPA has also outlined several specific problem areas where it sees a pressing need for innovation. These include developing systematic methods to understand the limits of what can be inferred from particular data sources, validating the connection between observed indicators and underlying psychological constructs, and designing models that can adapt over time without becoming obsolete.
The agency also highlights the need to integrate multiple psychological theories to better capture the interrelationships that define human behavior—from identity formation to collective decision-making.
Proposals will be funded under DARPA’s Advanced Research Concepts (ARC) framework, which is designed to give individual researchers the freedom to explore emerging, high-risk ideas.
Researchers hoping to take part must submit an abstract outlining their proposed approach. DARPA requires that submissions go beyond general overviews to define specific problems and offer quantitative benchmarks comparing their new methods to existing ones. A plan for validating predictions against real-world data is also essential.
Selected proposers will be invited to present their ideas in detail, with the potential for funding through a Research Other Transaction (OT) agreement—a flexible contract vehicle often used by DARPA to accelerate cutting-edge projects. However, funding is limited, and the agency warns that not all strong proposals will necessarily receive support.
The solicitation notice emphasizes that the Pentagon is keen to attract new voices, especially those with fresh ideas that could help crack the code of complex human systems.
The submission window for abstracts closes on June 30, 2025. However, DARPA encourages early submissions, noting that the opportunity could close sooner if funding runs out.
The solicitation notice emphasizes that the Pentagon is keen to attract new voices, especially those with fresh ideas that could help crack the code of complex human systems.
As society grapples with unprecedented challenges—from pandemics to information warfare to climate-driven migrations—the need for better tools to predict and understand collective human behavior has never been clearer. With MAGICS, DARPA is betting that new methods and paradigms have yet to be discovered to meet this challenge.
“MAGICS seeks innovative solutions that can better handle the challenges of systems that constantly change, interact with their environment, and respond to continuous self-observation,” DARPA writes. “Ultimately, the goal is to spur the development of novel, rigorous methods, and frameworks that advance our collective capacity to understand and predict human behavior with greater accuracy and nuance.”
Tim McMillan is a retired law enforcement executive, investigative reporter and co-founder of The Debrief. His writing typically focuses on defense, national security, the Intelligence Community and topics related to psychology. You can follow Tim on Twitter: @LtTimMcMillan. Tim can be reached by email: tim@thedebrief.org or through encrypted email: LtTimMcMillan@protonmail.com