Program to track social media for signs of depression, anxiety
Diana Inkpen, a professor of computer science at the University of Ottawa, is spearheading the development of a new Artificial Intelligence (AI) software designed to detect signs of mental illness online.
This technology uses machine learning algorithms to understand what people think by filtering texts, and examining online data on platforms such as Facebook, Instagram, and Twitter to find indicators of mental illness to be later analysed by experts in mental health.
To develop the program, Inkpen is working with a team of other computer scientists, along with psychiatrists, psychologists, master’s students, and her own computer science students. The team has been working on the research for two years now and are planning to have most of the program developed in the next year. Currently, the team is working to perfect the user experience aspect of the program.
Inkpen believes that people tend to turn towards social media to share their opinions and discuss personal stories, and by monitoring social media, the AI technology will be able to pick up on text or images that denote negative emotions by the user.
According to Kaitlin Keyes, a fourth-year English student at the U of O, “It would be cool if you were looking at a ‘triggering’ hashtags—it would give you an alert and ask you if you are okay. It would be cool if the app itself could help you get better. For example, giving tools to students such as breathing techniques or mental illness resources.”
Although doctors and other mental health experts are needed for the annotating of the data and advising what the program should look out for, Inkpen explains that researchers can use hospital records to examine symptoms that the program should currently look out for.
To guarantee the validity and reliability of this program, Inkpen explained that a well-trained program should be able to analyze the language of social media users to deter false statements. The researchers on this program are also working on perfecting this program to not miss out on important cases.
“The more accurate this technology is, the more it will help in our lives,” said Inkpen.
According to Inkpen, in-person mental health services are often full and have lengthy waiting times for parents to see a specialist and voice their concerns about their children’s mental health. Inkpen and her team hopes to be able to prevent and treat mental illnesses based on early warning signs.
Inkpen also believes that this technology would be effective in terms of cyberbullying detection or aggressions, based on language use and images. This tool could also help the authorities, by making it possible to guess the location of the people committing such behaviour.
In order to maximize the benefits to students and the university, this tool can also be made accessible by university administration and mental health counselors across campuses.
