How AI Supports Student Mental Health in Higher Education

How Technology Can Reinforce Students’ Mental Health

After observing a disconnect between the number of students who were struggling and how few actually sought assistance, Yusen Zhai, director of the Community Counseling Clinic at the University of Alabama at Birmingham, created a tool to predict which individuals may be more likely to experience anxiety or depression.

“I started wondering if we could use existing, common data to identify those who might be at risk before their situation worsens,” Zhai says. “The idea was to give mental health professionals and institutions a way to intervene earlier.”

Universities could potentially use the tool to analyze large data sets — such as the information all students provide at enrollment — to detect patterns that suggest groups sharing certain characteristics may develop anxiety or depression.

Similarly, counseling centers might assess a student’s medical history or other past treatment data to help tailor the care they provide, pinpointing risk factors that aren’t immediately visible, so mental health professionals can look for related signs and determine if a follow-up or referral would make sense.

Zhai has plans to further test the AI model and is examining ways to implement it.

“Some students are less likely to seek help for a range of reasons, including financial hardship, stigma, past experiences or not knowing what’s available,” he says. “This tool gives counselors and institutions another way to find and support those students, without waiting for them to come in on their own.”

Undergrads typically deal with a significant amount of stress, says Marla Johnson, technology entrepreneur-in-residence at the University of Arkansas at Little Rock, who believes AI in tandem with mental health specialists can allow universities to proactively address students’ needs and prevent crises from occurring.

In June, more than three dozen college and high school students attended a week-long AI & Mental Health Hackathon event planned by Johnson and held on the university’s campus.

“We wanted to show how AI could be applied to make a positive difference in the world and expose our students to what that process would look like,” she says.

Through courses from NVIDIA and other instructors, attendees learned about prompt engineering, AI ethics and other topics. Working in teams, they developed plans for tech solutions, such as software to assist a regional veterans’ healthcare provider with scheduling and prioritizing mental health service requests.

“Mental health is still somewhat stigmatized in Arkansas and the entire heartland,” Johnson says. “So, thinking of ways we can use technology to provide more access is very important.”

Maximizing the Impact Services Have on Students’ Lives

While studies have indicated AI has the potential to meaningfully contribute to mental health research and practice, schools may want to strike a balance between tech and human interaction, particularly considering AI models’ current communication capabilities.

Chatbots might actually validate harmful beliefs users express instead of questioning them, according to doctoral student Zainab Iftikhar, who is pursuing a Ph.D. in computer science at Brown University. The functionality tends to produce repetitive, formal-sounding responses — due to its propensity for perfect grammar and sentence structure — while failing to pick up on nuanced language and other emotional cues.

In 2024, Iftikhar conducted a study in which deep-learning models reviewed transcripts of human conversations and erroneously determined the interactions did not involve any empathy. The real-life participants, however, confirmed they had.

“There’s a lack of contextual understanding,” she says. “Unlike trained professionals, who can change their therapeutic modality based on a patient, these models just follow the prompt or instructions given to them. AI can help train peer supporters who are not psychiatrists in providing better care, but substituting all human support with AI would be more harmful than beneficial.”

AI technology could have other possible utilizations at colleges and universities;  answering basic questions about how to receive mental health services, for instance, or automating aspects of counseling center employees’ daily tasks.

“Especially in higher ed, we're definitely seeing more of the workforce using AI to streamline operations where possible,” Muscanell says. “It can be used as a support, so it frees up time for some of the staff.”

To encourage students to take advantage of any care-related tech tools the school offers, AI and mental health literacy initiatives can help highlight the positive outcomes they’re able to provide, according to Muscanell. She also suggests colleges and universities make sure students can easily access the solutions.

“They don’t want to look at seven different spots to find these resources,” Muscanell says. “Institutions need to make an effort to take a student-centered, unified approach;  possibly, for example, integrating these new tools into platforms students already heavily rely on, such as learning management systems. Then, they'll be more likely to use it.”