Revolutionizing the Role of Apps and Wearables for AI-Powered, Personalized Mental Health Care

Can AI-powered video games and smartwatches diagnose autism or prevent substance use? Meet Peter Washington, PhD, who is pioneering AI-driven diagnostics and treatments—making them more precise, personal, and accessible beyond the clinic.

Imagine a world where video games do more than just entertain: they serve as diagnostic tools for ADHD or autism. Or where your smartwatch doesn’t only count your steps but helps detect physiological signs of anxiety or substance cravings, then offers personalized mental health interventions.

This is the future Peter Washington, PhD, is working to make a reality. As a new assistant professor in UCSF’s Division of Clinical Informatics and Digital Transformation (DoC-IT), Peter is pioneering AI tools that turn everyday technology into powerful—and accessible—healthcare solutions. His research on consumer digital health informatics blends artificial intelligence (AI), consumer devices like wearables and smartphones, and human-computer interaction to develop AI tools that revolutionize diagnoses and treatments for mental health and neurological conditions.

“I’m really lucky to have the opportunity to guide consumer digital health informatics at a place like UCSF,” Peter said. “There’s direct access to a diverse population of patients who want and need these solutions as well as clinical collaborators who understand the potential of consumer digital health to improve patient lives. The ultimate goal is to translate this research into real-world practice, which is why I’m doing this work.”

Can AI and Video Games Improve Mental Health Diagnoses?

While AI models are excellent at detecting objective patterns in data, they often overlook the complexities of human behavior and individual variability. For example, an AI tool analyzing speech patterns might flag monotone speech as a sign of depression yet miss that the individual naturally speaks in a flat tone rather than having trouble feeling or expressing emotions.

To overcome this challenge, Peter’s “human-in-the-loop” approach keeps humans involved in the AI process by incorporating real-time input from doctors, caregivers, and patients. This helps the AI improve its predictions and adapt to the subjective elements of psychiatric and behavioral diagnoses, which often rely on personal experiences and symptoms that AI alone would be unable to accurately differentiate.

His current project, SocialPlai, is an AI-driven browser game that transforms mental health screening for ADHD and autism into an engaging, interactive gaming experience. The game analyzes players’ decision-making, social interactions, and facial expressions in real time to train AI models that distinguish between diagnoses. With plans to launch participant recruitment by late 2025, SocialPlai aims to make early screening more accessible, scalable, and engaging than traditional clinical assessments.

Looking ahead, Peter seeks to tackle big healthcare challenges, like eliminating high barriers to diagnosis, helping doctors work more efficiently, and reducing physician burnout. By exploring questions about the relationships and workflows among clinicians, patients, and AI, he aims to optimize innovations to aid in real-world clinical decision support. His next steps involve closely collaborating with all these groups to refine platforms like SocialPlai, ensuring AI insights are easy to understand and use in real-world clinical settings. He’s also diving into developing new digital tools that will allow patients to track important health data from home, making the diagnostic process simpler and more accessible for everyone.

AI That Adapts to You: Personalized Mental Health Interventions Through Wearables

You may have heard of breakthrough AI tools that analyze tissue samples for cancer diagnoses, automating tasks that traditionally required expert pathologists. While most of these rely on large datasets from thousands of patients, Peter is focused on developing AI that’s personalized for each individual. This approach is particularly important in that it helps reduce biases that can arise when AI is trained on historical data, ensuring more accurate and equitable diagnoses for all patients.

One example of Peter’s projects in this area focuses on how wearables can detect mental health events. By analyzing an individual’s unique biometric patterns, such as heart rate and skin conductance, these tools would identify or prevent cravings related to substance use disorders. Unlike traditional AI, which struggles with physiological variability, his approach continuously adapts to each person, providing early warnings and tailored interventions like digital peer support or cognitive behavioral therapy. With promising early results, this research could revolutionize addiction recovery and mental health care by making interventions more precise, proactive, and accessible.

Moving forward, Peter envisions taking this work even further by integrating these solutions into clinical practice at UCSF Health, developing best practices for their implementation, and embedding effective technologies into the health system’s infrastructure. He has his sights set on collaborating with teams at the Zuckerberg San Francisco General Hospital and Trauma Center, the UCSF Fontana Tobacco Center, and others to co-design digital therapeutics with both patients and behavioral health specialists. Projects like this would also focus on remotely monitoring cardiovascular disease or chronic and gestational high blood pressure.

By bridging consumer technology with clinical informatics, Peter aims to bring AI-driven solutions directly into people’s everyday lives, as well as make health care more precise, equitable, and effective.

Combining Creativity with a Dedication to Improve Lives

Peter’s research is driven by a deep commitment to transforming AI into a powerful tool to impact patient care. His dedication to this work is deeply personal, influenced by both his own experiences and the health challenges faced by his family and friends. He is also committed to mentoring the next generation of researchers and has championed projects driven by his trainees’ healthcare passions.

Before joining UCSF, Peter completed his PhD in bioengineering at Stanford University, MS in computer science at Stanford University, and BA in computer science at Rice University. He then spent over two years as an assistant professor at the University of Hawaiʻi at Mānoa. Now, with a move back to the Bay Area, he’s eager to collaborate with UCSF’s experts to bring his work into real-world patient care.

“Having the opportunity to make a meaningful impact in areas I care about by using my hobby of developing creative technologies is incredibly rewarding,” Peter said. “Knowing that this work can directly benefit others and contribute to better and more equitable patient outcomes makes it all the more fulfilling.”

Peter Washington, PhD

Assistant Professor, DoC-IT

__

We want to hear from you!
Inspired to support Peter’s work or discuss ways to collaborate on these efforts? See below for ways to connect with us.

About the UCSF Division of Clinical Informatics and Digital Transformation (DoC-IT)
DoC-IT serves as the academic home for applied clinical informatics researchers within the UCSF Department of Medicine. We also serve as a coordinating entity with key internal and external digital stakeholders across all UCSF mission areas, schools, departments, and divisions. Clinical informatics is approached as a multidisciplinary field that involves the use of technology by a broad spectrum of health professionals, patients, and other stakeholders.

Contact
Mika Rivera, Director of Communications
[email protected]

Follow Us
X (formerly Twitter): @UCSF_DoCIT
Bluesky: ucsfdocit.bsky.social
LinkedIn: www.linkedin.com/company/ucsf-doc-it