Is Smartphone Cannabis Impairment Detection Feasible?

The rise of the marijuana industry in the United States is upon us, and existing cannabis intoxication detection methods such as urine, blood, or saliva tests were not designed for cannabis and have significant limitations. Simply testing positive for the THC metabolite is not an indicator of impairment. However, a study published in the Drug and Alcohol Dependence, led by Steven's Institute of Technology assistant professor, Sang Won Bae reveals how smartphone technology may help bridge the gap.

With the use of the co-developed smartphone application, machine-learning models were used by Bae in research for detecting binge drinking. In addition, according to a press release, the algorithm could help law enforcement, employers and health professionals predict if a person is currently experiencing 'cannabis intoxication’, as well.

Is it truly as accurate as it is perceived? This testing method raises a conversation amongst the cannabis community, as well as, your average adult-use consumers and medical patients.

Detecting Cannabis "Highs" on Your Phone

The study was released in an issue of Drug and Alcohol Dependence in November 2021. Scientists evaluated the research for the feasibility of using smartphone sensor data to identify episodes of 'cannabis intoxication - in other words - being noticeably high.

Getting high is typically linked to slower response times, impaired driving, and other focus-intensive activities. In addition, existing cannabis detection measures, such as urine and blood, present their issues and would be time-consuming to use as an intervention.

As more states begin to legalize marijuana, many believe that a smartphone app could determine acute cannabis intoxication instead of general usage. The study's authors include faculty members from Steven's Institute of Technology, Carnegie Mellon University, Stanford University, University of Washington, and the University of Tokyo. They say that future research should investigate how algorithms like theirs would classify individuals who use cannabis less frequently.

Is it as simple as it seems? The research into AI technology and determining cannabis intoxication will be more of a complex challenge when put into action for the front lines for interventions and DUIs.

What Kind of Weight is There to this Newfound Technology?

Bae and her colleagues, including those at Carnegie Mellon University and Rutgers, discovered a combination of two sets of data that predicted cannabis intoxication with a 90% accuracy in a natural environment.

She created the AI technology to detect cannabis intoxication, which can potentially be applied to see the rise of risky behavior, leading to early interventions in everyday settings. Bae used over 100 features to detect whether each participant was intoxicated, including GPS, light, noise, and activity levels. Researchers also looked at factors such as the day of the week and time of the day in smartphone usage, while subjects self-reported being either "sober" or "high."

"It's important to give people the chance to change their behavior before something negative happens," Bae stated. "This study aims to predict human behavior as a way to support people while physically or cognitively impaired."

With typical cannabis testing, such as blood or urine tests, THC use can be detected from as much as one or two months ago. However, the smartphone app can be accurate with acute on-set effects.

Cannabis Sobriety is Difficult to Detect

Some properties are unique to the cannabis plant, but one that is true to just about any substance is tolerance. The level of tolerance depends on how often, what dose, and the potency it takes to create the desired effect. And it varies from person to person. Patients with medical marijuana cards, for example, may consume marijuana every day and seem unaffected, and quickly pass a field sobriety test.

But, beyond that, weed moves through the body in a very different way from alcohol. Marijuana and its specific psychoactive component leave the blood quickly, but it sticks around in the fat in the brain. Dr. Richard Clark is the director of medical toxicology at the University of California, San Diego. He believes that cannabis may even move from these tissues back into the blood days later in 'chronic' smokers.

The Future of Cannabis Intoxication Detection Lies in Technology

Although the science may still be developing solid and precise intoxication testing for cannabis. Some companies currently sell THC breathalyzers and several other tests for development but face the same challenges of disparity between blood levels and actual highs.

New technology from PredictMedix uses specialized multi-spectral cameras and complex AI to determine impairment factors. Another company, Zentrela, uses brain wave measurements to determine cannabis impairment.

There are many details, studies, and tests to be done before understanding how to properly test for cannabis intoxication, but the technology is catching up.