Michael Stevens, PhD, director of the Clinical Neuroscience & Development Laboratory at the Olin Neuropsychiatry Research Center, recommends waiting three to six hours after taking THC before driving. That wide range reflects the uncertainty researchers face in determining cannabis impairment.

Stevens and his team recently conducted a lab study using a high-tech driving simulator. They tracked 19 performance metrics over several hours and examined brain activity through MRI. Six metrics, roughly one-third, showed significant impairment following cannabis use. Drivers passed cars too closely, tailgated, and failed to make steering adjustments to keep vehicles centered in lanes.

Higher-potency cannabis led to worse driving performance than lower doses. However, THC levels were not a strong predictor of dangerous driving or when it would occur. Some participants' dangerous driving faded gradually while others' worsened before improving.

A concerning finding: about two-thirds of people said they'd be willing to drive even while admitting they felt high and after researchers measured they were objectively driving worse. A recent AAA survey found 35 percent of people believe they drive better when high.

Stevens notes some truth exists there. People can be more cautious when stoned, but cannabis impairs many cognitive abilities. Drivers who were high tended to drive too slowly, which disrupts traffic flow, encourages risky passing maneuvers, and causes rear-end collisions.

Determining THC impairment remains difficult. A 2021 analysis found a 15 percent increase in fatal car crashes following marijuana legalization, and a separate review linked legalization to increases in car accidents, injuries, or fatalities.

Some states (Illinois, Montana, Nevada, Ohio, Washington) set legal thresholds between 2 and 5 ng/mL THC in blood. Other states (California, Massachusetts, New York) use an effects-based approach looking for signs of impairment. Tech firms have raced to develop reliable roadside cannabis Breathalyzers, but none have solved the core problem: measuring THC is not the same as measuring impairment. THC sticks around in body fat, so frequent users have a reserve that skews test results.

Stevens recommends the safest approach: when in doubt, wait it out. Call an Uber, text a friend, or stay put.