Decoding the Future: The Outcomes of the Generative AI-Powered mHealth Think Tank
Last week, the research community witnessed a pivotal moment as the Invitation-Only Visioning Think Tank on Generative AI-Powered Interactive mHealth Interventions concluded at the University of Michigan Campus in Ann Arbor, MI. Hosted by the mDOT Center in conjunction the d3center (PI: I. Nahum-Shani – CP8), this dynamic and forward-thinking event successfully brought together an exclusive cohort of leaders to tackle the future of digital behavioral health. The high caliber of expertise and the foundational strategic roadmap developed promise to revolutionize how we approach personalized interventions.
An Exclusive Gathering of Cross-Domain Leaders
The think tank, held on Thursday, October 9, 2025, was highly exclusive and designed to converge the fields necessary for true innovation. The attendees comprised an elite group of experts from academia and industry. Domains represented included behavioral science, Artificial Intelligence (AI), mobile health (mHealth), and human-computer interaction (HCI). Participants included core mDOT and d3c Center Investigators, as well as invited researchers and industry experts from institutions like Harvard University, the University of Michigan, UIUC, UCLA, and companies such as Google and CareEvolution, Inc..
This breadth of knowledge allowed for truly interdisciplinary collaboration, focusing on how new advancements in Generative AI, particularly large language models (LLMs), could be combined with emerging wearable and sensor technologies.
The Core Discussions: Opportunities, Challenges, and Ethics
The primary goal of the collaboration was to envision how to develop and deploy the next generation of JITAIs (just-in-time adaptive interventions) supported by Generative AI. These models create powerful opportunities for intelligent, interactive health interventions that can generate high-quality content (text, audio, image, video) with fluency and contextual awareness.
Participants engaged in structured discussions aimed at achieving several key objectives:
Exploring New Frontiers: Discussions dove into how these tools can unlock new ways to design JITAIs that respond to individuals in the moment. For instance, abstracts highlighted deep dives into Generative AI-Powered Interactive Self Reflection Interventions and the application of Generative AI to produce multimedia intervention content for visual guidance in complex tasks.
Addressing Stress Management and Autonomous Agents: Researchers discussed leveraging LLMs for profound self-reflection to promote healthier behaviors, building upon pre-Generative AI era studies that successfully showed reduction in self-reported stress intensity and frequency through data visualization. Furthermore, a significant focus was placed on the path “Toward Autonomous mHealth Agents,” utilizing multimodal LLMs and foundation models to integrate world knowledge with sensor inputs, thus enabling new forms of reasoning over noisy data.
Achieving Holistic Intelligence: Discussions covered methods for expanding the state space of JITAIs, utilizing LLMs to help create adaptive intervention agents that have a more holistic understanding of participants.
Navigating Ethical and Technical Challenges: A major component of the discussions involved examining key barriers—technical, clinical, and societal—that must be overcome. Participants rigorously explored the ethical, privacy, and algorithmic challenges that arise when applying generative models to health contexts, striving to design systems that are transparent, equitable, and trustworthy.
The Aftermath: A Foundation for Transformative Potential
The conclusion of the Think Tank marks a crucial launch point. The closing session successfully summarized the outcomes and contributions, setting the stage for the community’s next steps.
The most significant immediate outcome was the development of a strategic roadmap. Participants collaboratively defined a shared vision for the future of Generative AI-powered JITAIs, identifying key priorities, research gaps, and implementation pathways intended to guide innovation over the next 3–5 years.
The potential generated by this focused, interdisciplinary forum is immense. It achieved its goal of offering inspiration and providing a foundation for profound and transformative collaboration. By outlining concrete next steps and fostering new partnerships, the Think Tank has built considerable momentum, strengthening the community committed to advancing the science and impact of AI-driven behavioral interventions. Ultimately, the outcomes of this meeting are expected to revolutionize behavioral health by shaping how Generative AI, wearables, and JITAIs interact.