Dartmouth is a key partner in the new Institute of Artificial Intelligence, supported by the National Science Foundation, and builds on the university's pioneering research into the development of digital treatments for the treatment of addiction and behavioral and mental disorders.
Dartmouth will join the “Youth Inspiration” Research Center at the Multicenter AI Institute for Interactions for AI Assistant or ARIA. Faculty and researchers based in the Geisel School of Medicine's Technical and Behavioral Health Center, as well as Cognitive Science's Computer Science Department and Program, will guide the implementation of AI on devices and wearable sensors that can provide personalized assessments and interventions to users.
Based at Brown University, ARIA is supported by a five-year $20 million NSF grant, and aims to develop next-generation AI assistants who can interpret people's unique behavioral needs to provide safe and effective feedback in real-time.
ARIA was announced as one of five National Institutes of Artificial Intelligence, one of five National Institutes of Artificial Intelligence announced through a $100 million investment by the NSF. The Institute is consistent with the White House AI Action Plan to ensure US global leadership in AI development.
“Dartmouth is the birthplace of AI, so we are particularly excited to bring years of expertise to this National Institute of AI and work with exceptional partner groups across the country.”
“The goal is to create science that informs you of best practices for developing AI-driven agents that can trust with people about mental health and drug use-related needs,” says Marsch.
In the first year of ARIA, Dartmouth researchers will lead a project focused on identifying and adapting to rapid changes in physiology, behavior and cognition in people with major depressive disorders that occur before the onset of serious symptoms.
Quote
As Dartmouth is the birthplace of AI, we are particularly excited to bring years of expertise to this National Institute of AI.
In the second year, the Dartmouth team will focus on substance use disorders, particularly opioids, by examining complex interactions between physiological conditions, environmental triggers, and neuroconnectivity patterns leading to drug use. These insights combine with sensors, neurocognitive, and neuroimaging data to generate real-time interventions that help prevent recurrence and support long-term recovery.
“Based on established and continuous work to integrate behavioral sensing, AI, and data from theoretical and cognitive neuroscience, we understand the complex interactions between physiological states, environmental factors, and neuroconnectivity patterns that contribute to mental health and substance use behavior,” says Marsch.
The Dartmouth faculty at the heart of Aria's work are:
- Lisa Marsch, professor of psychiatry and biomedical data science at Andrew G. Wallace, ensures that practical use is the focus of Aria's research, education and outreach activities. She will also tap on CTBH's extensive partnership network to coordinate work with ARIA patients, mental health professionals, treatment facilities and advocacy groups.
- Andrew Campbell, Albert Bradley 1915 Professor of Computer Science in the 3rd century. Campbell, director of CTBH's emerging technology and data analytics core, will oversee the integration of new technical and scientific ideas into the AI prototypes behind ARIA's virtual assistant.
- Nicholas Jacobson, an associate professor of biomedical data science, psychiatry and computer science, is a collaborator at ARIA. Jacobson leads AI and Mental Health: CTBH's Technology Guide Healthcare (AIM High) lab innovation and builds primarily from the development of Therabot's lab, the first fully-generated AI psychotherapy chatbot to take clinical trials.
- Stephen Frankland, an assistant professor of cognitive science, is an ARIA collaborator who coordinates basic cognitive science research and computational modeling conducted at Dartmouth. His mind, brain, and computational lab focuses on the computational mechanisms underlying the cognitive flexibility needed to build more adaptive and sympathetic AI systems.
The only center of excellence designated by the National Institute of Drug Abuse, focusing on Digital Health Tools and Therapeutics, CTBH is home to many projects notifying ARIA, and Marsch led the development of the first FDA fraudulent prescription intervention for addiction treatment.
She also led the first study known as D-Tect, which projected the clinical trajectory of people in drug therapy and used behavioral sensing to design more accurate and personalized options for recovery support.
Campbell has led several projects demonstrating the ability of mobile and wearable technologies to assess mental health and predict the onset of symptoms. The longest four-year student life study conducted so far demonstrated how to measure and understand mental health using apps on students' mobile phones.
Campbell and Jacobson worked together on an app called MoodCapture, which combines AI with facial recognition software to use a person's phone front camera to ensure the onset of depression. We also developed the first system to combine behavioral sensing and CHATGPT to provide personalized mental health interventions, and MindScape, a framework that allows mobile sensing data to be used through learning models that support AI.
“Dartmouth is particularly focused on developing basic sciences to understand, predict and improve mental health outcomes through real-world interactions between students and clinical populations across the country,” says Campbell. “Where we stand today, the science of AI assistants for mental and behavioral health is in the early stages, and Dartmouth's previous ongoing research is well placed to help advance this emerging field as part of the ARIA.”
Jacobson's Aim High Group also works to create digital biomarkers that predict and respond to discernible levels of mental health changes that are essential for next-generation AI systems, says Jacobson. His group also works from a user's perspective, including organic use of models such as ChatGPT for mental health support, and how to assess clinical knowledge and bias in AI.
“This study directly informs you of goals within the Aria,” says Jacobson. “Our work demonstrates the promise of generative AI, but highlights the scientific research needed to ensure that these systems are safe, reliable and truly personalized.”
ARIA's cognitive science department aims to bridge the gap between the strengths of large-scale language models and the strength of human intelligence, Frankland says.
“It remains unclear whether the models we interact with our daily models are rational in the most useful way for humans,” Frankland says. “By understanding both human and machine cognition, we aim to create a system that allows the world and its users to grasp more flexible and reliable systems. This goal is nothing more important than treating mental disorders.”
Aria is led by Ellie Pavlick, an associate professor of computer science at Brown. Pavlick is a keynote speaker for AI and Student Mental Health Symposium, sponsored by CTBH, and was held at Dartmouth on March 17th.
In addition to Brown and Dartmouth, Aria includes many major research institutes. Colby College. New York University; Carnegie Mellon University; University of California, Berkeley. University of California, San Diego. University of New Mexico. Santa Fe Institute. Data and Society, Civil Society Organizations in New York. ARIA researchers meet annually with different partners and other partners, which are held annually. Dartmouth is scheduled to hold its lab's fifth meeting in 2030.
ARIA will also prioritize training the next generation of AI scientists through educational opportunities and curriculum for middle and high school students and two-year college students. Partner institutions can provide mentoring to undergraduate and graduate students, postdocs associates, and visiting scholars who can work at multiple universities and organizations.
Aria is also dedicated to working with mental health professionals, roundtables, and engagement in university-level certificate programs aimed at educating mental health professionals, as well as educating practitioners and implementing ARIA innovations in a more efficient clinical setting.
