It's a scenario all weed smokers face at some point or another: You've enjoyed a touch of the green, hours have passed and the effects seem to have come and gone. You feel okay to drive, but you're just not sure about which side of sobriety you'd fall on if your level of intoxication were judged by a trained professional. Well, now there's an app for that.

Over the last year and a half, Massachusetts-based retired psychology professor Dr. Michael Milburn built an app-based sobriety test called DRUID that’s designed specifically with the cautious cannabis user in mind. Milburn hatched the idea over a bag of vapor while celebrating a friend’s purchase of a shiny, new Volcano. The recently retired professor, who’d spent 40 years “figuring out ways to measure things,” wondered, in his words, “Gee, how would you measure how stoned a person is?” Considering Milburn’s long career researching and developing “ways to measure things”—the intellectual artifacts include a book called Sexual Intelligence, which introduced Milburn’s scale for sexual intelligence—the idle academic was the perfect person to examine the methods by which cannabis intoxication is measured.
Milburn discovered that current tests for determining if a person is high—which vary from state to state—are flawed.
If you’re pulled over in Oregon and an officer thinks you’re stoned, you’ll be asked to perform a series of field-sobriety tests. These are the ones you’ve seen on Cops for decades: walk a line and pivot, follow the flashlight with your eyes, count backwards from 70 to 50, etc.
If the officer deems the driver unfit to operate a vehicle based on their performance of these tests, the driver is then referred to a drug recognition expert who administers additional assessments, as well as a breathalyzer to rule out alcohol and a urinalysis to detect and measure the presence of cannabis metabolites. Should cannabis metabolites be found in the driver’s urine, DUI charges are filed.
The problems with this procedure are several and severe: First off, field-sobriety tests are tailored to expose alcohol intoxication, and they’re notorious for producing false positives when it comes to weed. These tests also rely on the subjective judgement of a police officer.
Perhaps more controversial is the urinalysis, which detects the presence of THC-COOH—a substance that can linger in the human body for weeks after a person last ingested cannabis.
Sure, you might catch a stoned driver with these methods, but false positives are just as plausible.
DRUID, Dr. Milburn’s sobriety-test app for the stoner set, offers a better way.
“DRUID is an app that provides a general measure of impairment using neuropsychological testing,” says Milburn. At the intersection of video game and roadside-sobriety test, Milburn claims the methods are backed by “data that shows [the app’s] reliability and validity in predicting impairment.”
DRUID works like this: A user signs in and is given the option to take a two- or five-minute test, after which they’re prompted to choose from three modules that detect symptoms of cannabis intoxication, such as reaction time and decision-making, hand-eye coordination and time estimation.
One module, intended to gauge a user’s reaction and decision-making time, involves clicking on or avoiding shapes that rapidly appear and disappear on a user’s screen. Another module assesses intoxication by testing a user’s time-estimation abilities—asking the user to judge when a minute has passed while tapping on circles at random spatial and chronological intervals. A third module has users stand on one leg for 30 seconds while trying to keep perfectly still, and a hand-eye coordination section where impaired drivers track a moving dot with their finger while counting the number of shapes that flash on the screen.
A new user is advised to practice the tests a few times before setting their baseline scores.
Milburn recently lent his technology to a police academy in Randolph, Mass., where DRUID was employed during training sessions. DRUID, Milburn claims, was more reliable and accurate in detecting intoxication than traditional officer assessments.
In this way, DRUID removes the subjective element that is human judgement—opting instead for quantifiable, data-driven analysis to determine if a person is indeed too stoned to drive.
Milburn can see applications in the insurance industry—a possible future where drivers use DRUID before they get behind the wheel, submitting the results to their insurance provider to secure cheaper premiums. Personally, I’d be surprised if companies like Uber and Lyft don’t eventually adopt technologies similar to DRUID—both as a safe-ride guarantee and as a way to lower their insurance premiums.