“Anyone with a smartphone or computer can perform image evaluation tasks. It’s low-cost, scalable, and tamper-resistant,” Bari said.
The task of ranking photos may seem easy, she says. But it assesses an individual’s unique preference profile among 1.3 trillion possibilities, creating an incredibly powerful tool.
The system utilizes concepts well known in the world of economics, such as loss aversion, risk aversion, and the desire to insure against bad outcomes, and quantifies a set of variables that represent human judgment.
The system was able to identify the type of substance used (meth, opioid, or cannabis) with up to 82% accuracy and the severity of addiction with up to 84% accuracy. Statistical evaluation of judgment data revealed that participants with higher substance use disorder severity were more risk-seeking, less resilient to loss, had more approach behaviors, and less variability in preferences, providing insight into the behavioral profile of individuals with substance use disorders.
By directly predicting substance use disorder behaviors, this approach will enable assessment across a broader range of addictions, which can include behavioral addictions such as excessive social media use, gaming, or food consumption, Bari said.
Featured image at top: Researchers at the University of California have created a new AI to help clinicians diagnose substance use disorders. Photo/image photo
