UI professor using data to better understand teen mental health

Professor Patrick Fan's son, Winston, was sad and shaken. A fellow West High swimmer had taken his own life.

"It was devastating, and he experienced a lot of trauma from it," recalled Fan, two years after the tragedy. "He asked me what we could do to help people in similar situations."

It should be noted Fan is the Henry B. Tippie Excellence Chair in Business Analytics, not a psychiatrist, counselor, or medical professional. But he believed something could be done. After all, he had applied AI to improve cancer care, so turning it toward mental health felt like a natural extension.

The large language models that underpin AI's power work by culling through mass amounts of data to recognize patterns.

What if, Fan wondered, could AI be trained to detect signs of emerging depression, anxiety, or suicidal ideation in large datasets of social media posts from teenagers? After all, young people are often reluctant to talk to parents or professionals, but they freely share their feelings online, especially on platforms like Reddit.

"Many teens won't talk to a counselor or their parents, but they'll open up to strangers online," Fan said. "We're building AI tools that can listen to those conversations, identify signs of distress, and give counselors a better chance to reach those who are suffering in silence."

Fan and his team of grad students built algorithms that sift through thousands of anonymous Reddit posts where people talk honestly about loneliness, anxiety, or depression. From those patterns, they created a specialized "vocabulary map" tied to emotional signals and context. That lexicon allows AI to recognize not just when someone may be struggling, but also why-whether the cause is grief, family conflict, alienation and isolation, or something else.

High school students are part of this work too. Through the University of Iowa's Secondary Student Training Program, Winston and more than a dozen other teens annotated thousands of posts, helping train the AI model to recognize emotional patterns more accurately.

Fan is also collaborating with fellow Tippie faculty members Emily Campion and Amy Colbert because of their research expertise in behavioral psychology, particularly in the workplace.

As a business analytics professor, Fan said his goal is not to offer a medical diagnosis, rather to give counselors, psychologists, and researchers a data-driven way to identify early warning signs faster and more objectively than a human. Ultimately, he wants to put this tool in the hands of those who need it most.

"Our goal isn't to make money-it's to make this accessible," he said.

"We want high schools, universities, and clinics to use it freely so anyone who needs help can get it. We plan to work with the University of Iowa's mental health counseling center to pilot and evaluate the tool using real student data, anonymized, of course. We want other researchers to use it as a benchmark and improve upon our model. The long-term vision is to create a global, open-source research community focused on AI-driven mental health detection and support."

The work promises to make a potentially profound impact on many, but it's also been therapeutic for Winston, Fan said.

"He's annotated thousands of social media posts himself and says reading them has given him a better understanding of how much pain some people endure," Fan said. "It's helped him process his own trauma and given him a sense of purpose. He told me this is the most meaningful project he's ever worked on. It's been fulfilling to include him and strive toward a common goal together.

"My passion is to use technology to create positive social transformation, and this project is one step toward that."

Deep Dive

Patrick Fan and son Winston discuss AI for teen mental health on the Tippie Leads podcast.

Read more from the Researchers on a Mission series here:

Use AI ethically

Love your job

Root out fraud

This article appeared in the Spring 2026 issue of Tippie Magazine.