You can consume cannabis legally in your state, at home, in the evening, go to sleep, wake up the next morning, drive to work completely sober, get pulled over for a broken taillight, and be arrested for driving under the influence. This is not a hypothetical. It happens regularly in states with zero-tolerance or per se THC limits because THC metabolites persist in blood and urine long after impairment has ended.
Cannabis DUI law is the single most consequential legal gap in the legalization framework. States have legalized possession and consumption without solving the fundamental problem of how to determine whether someone is actually impaired — as opposed to merely having used cannabis at some point in the recent past.
The Alcohol Model Does Not Work
Alcohol DUI enforcement benefits from a nearly perfect biomarker. Blood alcohol concentration (BAC) correlates closely with impairment: at 0.08% BAC, virtually everyone shows measurable cognitive and motor deficits. BAC also clears predictably — roughly 0.015% per hour — so a BAC reading tells you approximately when the person was drinking and how impaired they are right now.
THC has no equivalent biomarker. THC blood concentration does not correlate reliably with impairment. Peak impairment from inhaled cannabis occurs within 15-30 minutes, when blood THC levels are already declining rapidly. A frequent user may have 5 ng/mL of THC in their blood while completely unimpaired, simply because THC accumulates in fat tissue and releases slowly. A first-time user might be significantly impaired at 2 ng/mL.
Use the interactive DUI law comparison tool below to check how your state defines cannabis impairment, what THC limits apply, what penalties you face, and how these laws compare to neighboring states.
Despite this scientific reality, most states have adopted one of two approaches borrowed from the alcohol model — neither of which works well for cannabis.
Per Se Laws: Guilty by Chemistry
Seventeen states have enacted “per se” THC limits — a set blood THC concentration above which a driver is legally impaired regardless of observed behavior. The most common threshold is 5 ng/mL of active THC (delta-9-THC) in whole blood. Colorado, Washington, and Montana use this standard.
The 5 ng/mL threshold was not derived from impairment studies. It was selected as a policy compromise — high enough to avoid catching most casual users the morning after, low enough to catch recently-intoxicated drivers in many cases. A 2020 review in Clinical Chemistry concluded that no blood THC concentration reliably distinguishes impaired from unimpaired drivers across the range of individual variation in THC metabolism.
The problem is most acute for daily medical cannabis patients and heavy recreational users. These individuals may carry baseline blood THC levels of 3-8 ng/mL without any recent consumption — simply from accumulated THC slowly releasing from fat stores. Under per se laws, these individuals are technically in violation every time they drive, regardless of their actual cognitive state.
Zero-Tolerance States: Any THC Is a DUI
Twelve states maintain zero-tolerance policies where any detectable amount of THC or THC metabolites (THC-COOH) in blood or urine constitutes impaired driving. This is the harshest approach and the most scientifically indefensible.
THC-COOH — the inactive metabolite that standard drug tests detect — can remain detectable in blood for 7-30 days after last use depending on frequency of use, body fat percentage, and metabolic rate. A patient who uses medical cannabis for chronic pain and has not consumed in a week can still test positive for THC-COOH and face DUI charges under zero-tolerance statutes.
Arizona’s zero-tolerance law survived a legal challenge in 2014 (Dobson v. McClennen), but the state Supreme Court narrowed it to apply only to active THC compounds, not inactive metabolites. Other states have not made this distinction, leaving drivers exposed to prosecution based on metabolites that have zero pharmacological activity.
Effect-Based Standards: The Better Approach
A growing number of states use effect-based standards — requiring prosecution to demonstrate actual impairment through observed behavior, field sobriety testing, and Drug Recognition Expert (DRE) evaluation rather than relying solely on a chemical threshold.
Drug Recognition Experts are law enforcement officers who complete a 72-hour training program to identify impairment from specific drug categories through a 12-step evaluation protocol. The protocol includes pupil examination, vital signs, muscle tone assessment, and divided attention tests.
The DRE program has legitimate science behind it, but also significant limitations. DRE evaluations are subjective — they depend on the officer’s training, experience, and judgment. False positive rates in validation studies range from 5% to 20% depending on the drug category. And the 12-step protocol was designed primarily for opioids, depressants, and stimulants — its sensitivity to cannabis-specific impairment patterns is lower than for other substances.
The Technology Gap
The lack of a reliable cannabis breathalyzer has stalled progress on science-based enforcement. Several companies — most notably Hound Labs and SannTek — have developed devices that detect THC in breath, which correlates with recent use (within 2-3 hours) better than blood or urine testing. THC is only present in breath for a narrow window after inhalation, making it a better indicator of recent consumption than blood testing.
However, breath THC detection still does not measure impairment. A breath test can confirm that someone consumed cannabis recently. It cannot tell you whether that person is impaired. The gap between “recent consumption” and “impairment” is smaller for breath testing than for blood testing, but it still exists.
The ideal solution — an objective, real-time measure of cannabis-related driving impairment — does not yet exist. Some researchers are pursuing tablet-based cognitive testing batteries that measure reaction time, divided attention, and executive function at roadside. Others are investigating eye-tracking technology that can detect specific oculomotor patterns associated with THC impairment. None of these approaches are ready for deployment.
Penalties: The High Stakes
Cannabis DUI penalties in most states mirror alcohol DUI penalties — which means they are severe. First-offense penalties typically include license suspension (90 days to 1 year), fines ($500-$10,000), possible jail time (up to 6 months), mandatory substance abuse education, and probation. Second and subsequent offenses escalate to felony charges in many states.
These penalties apply even in states where cannabis is fully legal for adult use. The jarring reality: you can legally purchase cannabis at a state-licensed dispensary, legally consume it at home, and face a felony charge the next day based on metabolites in your blood that have nothing to do with your current cognitive state.
What Needs to Change
The National Highway Traffic Safety Administration (NHTSA) has repeatedly acknowledged that no scientific consensus exists for a THC impairment threshold comparable to the 0.08% BAC standard for alcohol. Despite this, most states continue to enforce per se or zero-tolerance standards that criminalize cannabis metabolites rather than impairment.
The path forward requires three developments: validated roadside impairment testing technology, abandonment of per se THC limits in favor of effect-based standards, and reform of zero-tolerance metabolite laws that criminalize past consumption rather than current impairment. Until all three happen, cannabis DUI law will remain what it is today — a legal minefield where the rules bear little relationship to the science.