Fascination of human behaviour guides postdoctoral fellow's career path

Tamara Vineberg - 26 February 2020

James Benoit helped to develop a chatbot as a postdoctoral fellow in the Department of Pediatrics.
Postdoctoral fellow James Benoit is fascinated by human behaviour. He’s been approaching the topic with different lenses that eventually led him to complete a three-month postdoctoral fellowship in the Department of Pediatrics, focused on using artificial intelligence (AI) to help vulnerable populations.

“I’m really interested in understanding human behaviour. My interest has gradually shifted and integrated more technology over the years to the point where now instead of looking at the basic science of behaviour, I am now looking at how I can help vulnerable populations,” says Benoit.

Benoit began his study of behaviour with a bachelor of science degree in neuroscience at the University of British Columbia where he examined the phenomenon of cognitive effort in rats. A master’s in applied ethics switched his focus to the moral status of violent video games and their effect on children.

During a PhD in psychiatry at the University of Alberta, he became interested in youth who display high risk behaviours and how their brains are different than youth who are not high risk. “Halfway through the PhD, I actually changed streams quite a bit. I started to work more and more with artificial intelligence and machine learning. We started to collaborate with a pharmaceutical company developing software that would allow us to figure out who would respond to antidepressant medication and who would not,” says Benoit. “It turns out when you apply AI to determine who will respond to treatment or not, you can get quite a significant increase in the results without waiting to see who will respond to an antidepressant. We determined who will respond to treatment for one specific antidepressant.”

At the close of his PhD, Benoit's supervisor suggested he look into an intriguing project in the Department of Pediatrics.

The idea is to build a chatbot — a computer program that can simulate a human conversation through text or voice — to help families living with neurodevelopmental disabilities find the best information and resources online. “We saw this opportunity in the lab to take what I’ve done in AI and healthcare and apply it in a slightly different way,” he explains. Benoit’s role was to participate in the coding and development of the chatbot. He also coordinated the interactions with social science, engineering and health specialists.

The chatbot allows parents to ask questions about their children’s behaviour and the chatbot will respond by asking more questions to narrow down the online resources that will possibly help. This allows parents to ask questions in a natural way and the chatbot will pick certain keywords as clues for determining appropriate information.

“Initially, we’re concerned that the resources we are recommending are the right resources for the question being asked. We want to ensure that parents are getting a good experience when they are using the chatbot. We want parents to be matched with high quality information that matches their search intent, and not frustrated by being given spurious results,” explains Benoit.

The experience gained during even a brief postdoc, such as this one, will add to Benoit’s portfolio as he pursues an academic career. His next step on this path is as a research fellow at Harvard University Medical School through the digital psychiatry program.