Driving under the influence of THC is illegal everywhere, full stop. What is far less straightforward, and what causes genuine confusion even for people who would never knowingly drive impaired, is how that is actually measured and enforced, and whether the legal thresholds in use actually correspond to impairment at all.
The Three Legal Approaches
US states take one of three broad approaches to THC and driving. Per se laws set a specific THC concentration in blood (commonly 2 or 5 nanograms per millilitre, ng/mL) that is treated as automatic, prima facie evidence of impairment, regardless of whether the driver shows any actual signs of impairment. Currently, six US states use this approach. Zero-tolerance laws make it illegal to drive with any detectable amount of THC or its metabolites in your system, regardless of concentration or impairment. Twelve US states use this approach. Impairment-based laws, used in the remaining states, require prosecutors to demonstrate actual impairment through evidence such as field sobriety tests, officer observations, and behaviour, without relying on a specific THC concentration as automatic proof.
Why THC Is Different from Alcohol for This Purpose
The per se approach works reasonably well for alcohol because blood alcohol concentration correlates fairly consistently with impairment across individuals, and alcohol is metabolised and cleared from the body relatively quickly and predictably. THC behaves very differently. THC is fat-soluble and can remain detectable in blood, and especially in fat tissue, for days or even weeks after use, particularly in regular users, long after any impairment has resolved. This means a THC concentration that might indicate recent, possibly impairing use in an occasional user could simply reflect residual THC from use several days earlier in a regular user who is not impaired at all.
What the November 2025 Research Found
A study published in Clinical Chemistry in November 2025, conducted by researchers at UC San Diego’s Center for Medicinal Cannabis Research, directly tested how well per se THC limits correspond to actual driving impairment. The study involved 190 regular cannabis users who were instructed to abstain from cannabis for at least 48 hours before testing. Despite this abstinence period, the results were striking: 43% of participants still exceeded zero-tolerance thresholds (≥0.5 ng/mL) at baseline, after 48 hours of abstinence. 24% had baseline THC concentrations at or above the 2 ng/mL per se cutoff used by some states. 5.3% were at or above the 5 ng/mL per se cutoff used by other states. The maximum baseline concentration observed, after 48 hours of abstinence, was 16.2 ng/mL, more than three times the higher per se limit.
Critically, the study also assessed driving performance using a driving simulator, and found that participants with these elevated baseline THC concentrations performed no worse on the simulator than participants below the per se cutoffs. The researchers concluded that current per se THC blood limit laws lack scientific credibility as face-value evidence of impairment, and that more work is needed to identify drivers who are genuinely impaired by cannabis and unsafe to drive.
What Earlier Research from AAA Found
This is not a new concern; the AAA Foundation for Traffic Safety had previously evaluated whether a scientifically supportable per se threshold for THC could be established, similarly concluding it could not. Their analysis found that a 5 ng/mL threshold would miss approximately 70% of cannabis-impaired drivers, who tested below that level, while potentially still penalising sober drivers who had used cannabis hours or days earlier and retained detectable but non-impairing THC levels. The fundamental challenge identified across this research is that, unlike alcohol, there is currently no established THC concentration that reliably predicts impairment in the way blood alcohol concentration does.
What This Means in Practice
If you live in a state with a per se or zero-tolerance law and use cannabis regularly, even for entirely legal purposes (medical cannabis, or recreational use in a legal state where you are not driving), you may have detectable THC in your system for an extended period, potentially days, even if you have not used recently and are not impaired. This creates genuine legal exposure that does not track with actual driving safety, according to the research discussed above. If you live in a state with an impairment-based approach, the legal standard focuses on actual signs of impairment rather than a specific blood concentration, which research suggests is a more scientifically defensible approach, though it relies more heavily on subjective officer assessment and field sobriety tests, which have their own limitations.
Practical Guidance Regardless of Jurisdiction
Regardless of the legal framework where you live, the practical safety guidance is straightforward and does not depend on navigating the legal nuances: do not drive while experiencing any of the acute effects of THC discussed in our piece on THC and the brain, including altered perception, impaired coordination, or significant changes in reaction time. The acute window of impairment from smoking or vaping is generally a few hours; for edibles, given their longer duration (six to ten hours or more, as discussed in our edibles piece), the window during which driving should be avoided is correspondingly longer. If you are a regular cannabis user in a per se or zero-tolerance state, being aware that detectable THC can persist well beyond any subjective effects is relevant to understanding your legal risk profile, even if it does not change the practical safety guidance of not driving while actually impaired.
International Approaches
Some countries have taken different approaches to calibrating THC limits. Norway, for example, has established THC blood concentration levels intended to correspond to specific levels of impairment comparable to defined blood alcohol concentrations, an attempt to create a more graduated, severity-based system rather than a single cutoff. Whether this approach is more scientifically defensible than the single-threshold systems common in the US remains a subject of ongoing research and policy discussion, but it represents a different conceptual approach to the same underlying challenge: THC’s pharmacokinetics do not map onto blood concentration in the same predictable way alcohol’s do.
Frequently Asked Questions
How long does THC stay in your blood after using cannabis?
This varies significantly based on frequency of use, amount, and individual factors, but research has found that regular users can have detectable THC in blood at levels exceeding common legal per se limits even after 48 hours of abstinence, with some individuals showing levels several times higher than the 5 ng/mL threshold used in some states. This is notably different from how quickly alcohol clears from the blood, and is the core reason blood THC concentration is considered an unreliable indicator of recent use or impairment for regular users.
Can I get a DUI for cannabis even if I’m not high?
In states with zero-tolerance or per se laws, yes, this is possible in principle. Research has found that a substantial proportion of regular cannabis users have detectable THC, sometimes above per se thresholds, even after 48 hours without use and while showing no impairment on driving simulator tests. In states with impairment-based laws, the focus is on evidence of actual impairment rather than a specific blood concentration, which would not capture this same scenario in the same way, though field sobriety tests and officer observations have their own evidentiary considerations.
Is there a breathalyzer for THC like there is for alcohol?
THC breathalyzer technology has been in development and limited testing, but is not in widespread standard use comparable to alcohol breathalyzers. Most current THC testing for driving enforcement relies on blood or oral fluid tests, which have the limitations discussed in this article regarding the relationship between THC concentration and actual impairment. The absence of a reliable, real-time impairment measure for THC, comparable to how a breathalyzer reading correlates with alcohol impairment, remains a significant gap in cannabis-impaired driving enforcement.
Does CBD affect THC driving tests?
CBD itself is not what THC driving tests measure. However, if you use full spectrum CBD products containing trace THC regularly, this could contribute to detectable THC levels in blood, similar to the broader discussion in our piece on CBD and drug testing. In a state with zero-tolerance or low per se limits, this could theoretically create the same kind of legal exposure discussed in this article, even though the CBD product itself was not intended to produce intoxication or impairment.
How long should I wait to drive after using an edible?
Given that edibles produce effects that can last 6 to 10 hours or longer, as discussed in our edibles piece, and that onset itself can take up to 2 hours, the window during which driving should be avoided after consuming an edible is considerably longer than for smoking or vaping. There is no universally agreed specific waiting period, but waiting until you are confident the edible’s effects, including any lingering effects into the following hours, have fully resolved is the relevant safety consideration, independent of what your specific state’s legal THC threshold happens to be.
Why hasn’t the per se limit system been fixed if research shows it doesn’t work?
Changing established legal frameworks requires legislative action, and per se limits, while criticised by researchers, do offer law enforcement and courts a seemingly objective, easy-to-apply standard, which has practical appeal even if the underlying science is contested. Developing and validating alternative approaches, such as standardised impairment testing protocols or technology-based real-time impairment measures, is an active area of research, but translating research findings into changed state laws is a separate and often slower process than the research itself.
