‘AI in mental health’ is the theme of a computing seminar
UFRJ professor Paulo Mann was the guest speaker on Tuesday (31)
Artificial intelligence (AI) is already a reality in society and in contemporary scientific literature. With this in mind, IMPA Tech’s scientific seminar this Tuesday (31), addressed how AI can support mental health, especially in the context of digital hyperconnectivity. Paulo Mann, a professor at the UFRJ (Federal University of Rio de Janeiro) Computing Institute, was the guest speaker at the presentation, which brought together students from the three bachelor’s degree classes in Mathematics of Technology and Innovation.
Mann surveyed recent articles, news from international outlets and presented studies from his research group, which investigates how computational methods can be used in the diagnosis, treatment and risk prediction of mental health issues. “We talk about AI all the time, but how can we use the tools to improve the diagnosis of people with depression? By analyzing data, we can have mechanisms for tracking symptoms, helping patients to seek professional help,” he said.
In his research, the professor used language models that cross-reference textual data and images to understand the behavior of users on social networks. “We take the image from Instagram, the associated caption, put all the posts in chronological order and send it to the model. The model tries to do this screening, analyzing whether or not the individual may be in a depressive episode,” he explained.
Interestingly, the results also pointed to a curious fact about technological society: the performance of happiness on digital platforms. “We noticed that the people with the highest levels of intensity of chance of depression, at the most severe levels, were those with the most beautiful and happy posts on social networks,” said Mann.
Although AI is an alternative for tracking behavior and supporting the treatment of mental health issues, the chat also warned of ethical and digital security issues. “When working with AI, we have impacts and biases on people with depressive illness. The data used to train models such as ChatGPT, for example, directly affects the models’ communication. Another issue is the protection of the data used, especially in the context of mental health. In addition, we can’t help but think about human connection and cognition. The feeling I get is that sometimes we’re dumbing down, so much so that we delegate tasks to AI tools,” he said.
For student Gabriel Falcão, who is studying Data Science, the topic was very relevant. “AIs are increasingly present in our daily lives and are changing the way we live. Studying the social impacts of these tools is a natural path and there will certainly be many research topics on the subject,” he said.
Jonas Araújo, from the Physics emphasis, highlighted the possibility of training in AI within the IMPA Tech curriculum. “Everything involving artificial intelligence is very interesting. After their popularization, everything has changed and is increasingly interdisciplinary. The lecture inspired us to study more about the subject and try applications in other areas of knowledge. I want to take elective courses to relate new topics,” he said.
Professor Mann also holds a doctorate (2023) and a master’s degree (2019) from the Institute of Computing at the Fluminense Federal University (UFF) and has participated in several research projects in the areas of AI, machine learning and computational social science. His work spans several fields, including AI applications in mental health, financial time series analysis and natural language processing (NLP).
