Russia’s invasion of Ukraine has been going on for almost five years now. Among other consequences, the war has taken a toll on the mental health of a whole generation of Ukrainian youth who are growing up in a warzone.
To ameliorate this problem, researchers from Brown’s School of Public Health are turning to a chatbot.
In collaboration with the National Academy of Educational Sciences of Ukraine and the Institute of AI Problems for Ukraine, Brown researchers from the Global Mental Health and Flourishing Lab are in the process of creating Chat for Hope — an AI platform to provide mental health support to Ukrainian youth — which will be implemented in the coming years.
Shufang Sun, associate professor of behavioral and social sciences, associate professor of psychiatry and human behavior and associate director of the mindfulness center, first turned her research focus to Ukraine in around 2023.
Collaborating with researchers in Ukraine during Russia’s war has posed specific challenges, especially as concerns of safety fluctuated, Oleksii Shebanov, a staff research assistant in the lab, said. This was especially felt when a Russian missile hit the Ukrainian academy where some of the project’s Ukrainian co-researchers were based, killing two and injuring 13 in the institution.
“That was a very heartbreaking moment,” Shebanov said. “Every morning I’m doing a little security review for (our) lab manager to make sure our collaborators are alive and can work.” Limited internet connection, electricity and water supply add to the challenges that Chat for Hope must overcome, said Shebanov.
Working to improve adolescent mental health is especially important because of how it impacts later life mental health outcomes, Sun said.
Audrey Tyrka, chair of the Department of Psychiatry and Human Behavior, who was not affiliated with the study, wrote in an email to The Herald that “when danger and fear is high and support is limited, youth can develop severe anxiety, depression, post-traumatic stress disorder and risky behaviors such as substance misuse and suicidal behaviors.”
The current limited available support for mental health concerns was a driving force behind the development of Chat for Hope.
“There’s just not enough mental health specialists around the world, including in the U.S,” Sun said. “There’s been just such a shortage of mental health professionals, which means that most people who need help don’t get help, and the gap is particularly wide for low-middle-income country settings.”
Because of the shortage of school psychology professionals, the Chat for Hope team seeks to “tap into” the school system in Ukraine. Their hope is to leverage the network of teachers and existing professionals to promote the use of digital mental health care and screening.
To provide this care, the final Chat for Hope product will be a chatbot with which youth can engage. According to Sun, the chatbot will have a “deterministic style and conversation flow,” meaning that all possible answers to inputs are already pre-determined ahead of time.
“It’s more boring for the user, compared to ChatGPT, for example, where you can’t predict the answers, but it’s also safer because we know what each answer will be about, and it’s less addictive compared to the commercial AI models on the market,” Sun said.
Since Chat for Hope will primarily serve youth, the chatbot will incorporate many safety features such as a direct online connection to a real person if a user is in immediate danger, according to Sun.
A lot of the work in developing Chat for Hope has come from communication with youth in the warzone that have been “affected personally” and using their input when creating the chatbot.
Noah Philip, a professor of psychiatry and human behavior who was not affiliated with the study, wrote in a message to The Herald that it is important the Chat for Hope chatbot is diligently tested before it is rolled out.
“The potential for digital mental health interventions is substantial, yet demonstrating meaningful clinical benefit through rigorous research is essential before these approaches can be widely implemented,” Philip wrote. “The last thing people need in a conflict zone is something that doesn’t help, or even worse causes harm by delaying other evidence-based interventions.”
But if it does demonstrate “meaningful clinical benefits,” the chatbot could have an impact in conflict-affected regions across the world, Philip added.
The chatbot will first be tested via a randomized control trial on 480 Ukrainian youth. “Half of them will receive the specific intervention, and half of them will receive the screening and referral resources, but not the treatment that’s built within the app,” Sun said. The Chat for Hope team will then perform an evaluation and two follow up appointments.
“The last thing people need in a conflict zone is something that doesn’t help, or even worse causes harm by delaying other evidence-based interventions,” Philip wrote.
Shebanov said that it would be “amazing” if Chat for Hope would allow youth to “feel empowered (and) inspired to be resilient, to stand against different challenges that there may be, particularly related to war.”
Alice Xie is a section editor for Science and Research from Los Angeles, California. She studies Applied Mathematics and Biology, and enjoys reading gut wrenching literature in her free time.
