New AI Model Detects Mental Disorders Based on Web Posts

At the 20th International Conference on Web Intelligence and Intelligent Agent Technology, Dartmouth researchers unveiled an artificial intelligence model for detecting mental disorders using conversations on Reddit.

The research is “part of an emerging wave of screening tools that use computers to analyze social media posts and gain an insight into people’s mental states” , reports TechXplore. In a nutshell, the “AI model detects mental disorders based on web posts.’

One of the paper’s co-authors says “social media offers an easy way to tap into people’s behaviors.” Another benefit to using it is that the data is not only public, but was voluntarily published.

The new AI model is described as the best in its class because it is able to focus on emotions rather than the specific content of the media texts being analyzed. Technology, surveillance and big tech have always walked a fine line between legal and illegal.

The co-author says there are many reasons why people with mental health disorders may not want to seek help. Reasons such as stigma, high costs, and lack of access to services. But, the co-author adds that it’s possible with some prompting, those individuals will seek some help.

That’s where the issue can go down a slippery slope. The researchers “trained their model to label the emotions expressed in users’ posts and formed “a matrix that would show how likely it was that a user went from any one state to another, such as from anger to a neutral state of no emotion.”

Each emotional disorder has its own patterns of emotional transitions. Therefore, an emotional “fingerprint” for a user is created and compared to emotional disorders.