Australian scientists from the Royal Melbourne University of Technology have developed an artificial intelligence (AI) model that can accurately determine a person’s risk of developing a video game addiction. The algorithm uses data about players’ relationships with their virtual avatars, their ages, and the time they spend playing the game. The study was published in the scientific journal magazine Journal of Behavioral Addictions (JBA).
Gaming disorder is a mental illness recognized by the World Health Organization (WHO). It is characterized by preoccupation with video games to the detriment of other activities, impairment of self-control, and reluctance to stop gaming despite negative consequences.
To be diagnosed, symptoms must be severe enough to significantly disrupt personal, family, social, occupational, and other important areas of life. This usually requires at least 12 months of patient observation.
565 players between the ages of 12 and 68 participated in the research. About half of the participants were men. They reported up to 30 years of gaming experience, with an average duration of 5.6 years. They had also been using social media for an average of seven years and spent approximately three hours a day on these platforms. Of these, 55% were working full time, 36% had a bachelor’s degree, and 30% were single.
The authors of the experiment tested the participants’ condition twice, with six-month intervals between tests. Participants completed a diagnostic assessment of gaming disorder and an assessment of their connection with their in-game avatars. Second, they include degrees of identification with one’s character (e.g., “Both I and my character are the same”), immersion (e.g., “Sometimes I only think about my character when I am not acting”), and compensation (e.g., “I prefer to be like my character”).
At the beginning of the study, just under 20% of participants were determined to be at risk for gaming disorder. The scientists then split the dataset into two parts: 80% to train the AI models and 20% to test their prediction performance.
Once trained, the AI models were able to accurately identify participants at risk for gaming disorder based on user-avatar connection score, age, and gaming duration.
Parents of a former school child in the United States sent Developers of popular online games were sued over the addiction their children developed.