Artificial Intelligence is Helping Scientists Detect Loneliness

artificial intelligence

The start of the 21st century has had scientists witnessing an increase in loneliness amongst the population. Several studies in the psychological and sociological fields over the last decade have reported that large sectors of the population are incredibly lonely. And this very loneliness majorly impacts the overall happiness quotient of demography- with increasing suicide rates, depression rates, and substance abuse.

The Coronavirus pandemic has simply led to a growth in this trend, as people are unable to intermingle- something that they could have done in the past to get rid of loneliness. Even amongst those who experience loneliness, older adults and marginalized groups lead the race. According to a study conducted by MedicalXpress in collaboration with UC San Diego, the percentage of loneliness was around 85% in communities with senior housing.

Since the problem of loneliness seems to have no conventional solutions, scientists are looking at Artificial Intelligence as a way of solving this international crisis. 

How is Artificial Intelligence or NLP useful for Loneliness?

But before they actually determine any solution to this problem, scientists have to get a proper view of the situation. This is where they hit a major snafu, as most of the methods through which they collect data lack certain aspects. In many situations, self-reporting would only be used as an option if the individual finds themselves in the extreme cases of loneliness. When asked, most individuals won’t tell people how lonely they are, due to social constraints. This is where artificial intelligence or Natural Language Processing comes into play. With the methods of NLP, scientists try to analyze the original, authentic way with which people try using their language- something that would lead to a much more honest, and less biased representation of the loneliness of people. 

The study’s lead author Ellen Lee found their target group as people between the ages of 66 to 94- with a sample size of 80 participants. In this study, the participants were advised to answer to the best of their abilities without actually thinking much about them. The key wasn’t eloquence, but authenticity. Interestingly, the authors weren’t just having a question-answer session with the participants- the NLP they used allowed them to have long-form interviews- which then helped them find indications of loneliness. 

The Findings From the Artificial Intelligence Study

The findings stated that people who were lonely had several key differences in their interactions than people who were not lonely. Their answers were longer, and much sadder when asked about loneliness in itself. Gender too played a big role, as men were not as likely as women to admit being lonely. Also, men were more expressive about fear and joy than women. 

Artificial Intelligence thus used by the scientists accurately detected/predicted qualitative loneliness by 94%. Currently, the project has been geared towards using NLP methods to find out if there is any correlation between loneliness, and wisdom- especially in older adults. For, one thing had been proven, Artificial Intelligence could detect loneliness through speech pattern analysis. All that is left now is to take it even further.