New research into the link between America’s falling alcohol use and its rising marijuana consumption finds a road safety picture that is improving on one front and becoming harder to read on another. In 2023, 62% of Americans drank alcohol. By 2025, that figure had fallen to 54% — the lowest rate recorded in Gallup’s near 90-year polling history. At the same time, around 62 million Americans aged 12 and older reported using cannabis in the previous year. As one habit declines, another is growing. The question is what that swap means for the people sharing the road with those making it.
The answer, according to crash data, is cautiously positive — for now. NHTSA figures covering 2022 to 2023 show that total fatalities involving an impaired driver fell from 1,366 to 1,113. Thirty-three states recorded a decline in impaired driving deaths during that period. Thirteen saw an increase. Early projected estimates for 2024 pointed to further improvement, with total fatal motor vehicle crashes falling nearly 4% to 39,345 — the first time since the pandemic that annual road deaths dropped below 40,000, even as Americans drove more miles than the year before.
But the data also surfaces a problem that is growing at the same pace as the good news. Cannabis is now legal for recreational use in 25 states and Washington D.C. Its tourist market was valued at $10.23 billion in 2023 and is projected to reach $23.73 billion by 2030. More people are using it, more states are permitting it, and more drivers are getting behind the wheel after consuming it — often without a clear idea of whether they should.
Why Americans Are Putting Down the Drink
The decline in alcohol consumption is not a gentle drift. It has been sharp, generational, and accelerating. The drop has been steepest among women, down 11 percentage points since 2023 to 51%. Among men the fall is 5 points, to 57%. Non-Hispanic White adults have seen an 11-point drop. Republicans have seen the sharpest fall of any political group — down 19 points to 46%. Democrat numbers have held steady at around 61%.
Even among those who still drink, habits have changed. A record low 24% of drinkers said they had consumed alcohol in the previous 24 hours. The average number of drinks consumed in the past week was 2.8 — Gallup’s lowest figure for this measure since 1996, down from 3.8 in 2024 and a consistent 4 per week across the several years before that. Forty percent of drinkers said it had been over a week since their last drink, the highest proportion since 2000.
The reasons are generational as much as behavioral. A December 2025 survey of 2,000 adults found that 61% of Americans are losing interest in alcohol. For Generation Z, that figure rises to 63%, and the shift begins early: the average age at which Gen Z respondents said their interest in drinking started to fade was 23. For Baby Boomers, the equivalent age was 44. Health was cited as the top reason by 4 in 10 respondents, followed by a dislike of hangovers at 36%. The cultural signals are shifting too — 18% of survey respondents said drinking alcohol is simply out of fashion in 2026.
What Is Replacing It — And What That Means Behind the Wheel
THC drinks were named as one of the things that is actively in fashion by 18% of the same survey respondents. Nearly half of Americans — 48% — believe THC products should carry the same social acceptance as alcohol. Among Millennials that figure jumps to 60%. When asked to choose between methods, 33% of respondents said they would reach for a THC beverage over smoking marijuana at 28%. The product form is shifting, the consumer base is broadening, and the age profile of the typical cannabis user is getting older and more mainstream.
That is where the road safety picture becomes harder to parse. NHTSA survey data shows that 12.6% of weekend nighttime drivers tested positive for marijuana. A survey of 990 drivers in states where recreational cannabis is legal found that 37% use it daily. When asked how long they typically wait before driving after using cannabis, 35.5% said between one and four hours — a window that falls well short of the minimum recommended by states such as Washington (five hours for inhaled cannabis) and Colorado (six to eight hours depending on dosage and method).
Six percent of those surveyed admitted to using cannabis while actively driving. And of the drivers who received a traffic ticket in the previous year, 41% admitted they were under the influence of cannabis at the time. These are not fringe behaviors. They describe a significant proportion of drivers in legalized states who are routinely making risk judgments about their own impairment that the data suggests they are not well placed to make.
The Confidence Gap: What Drivers Think They Know
Perhaps the most striking finding in the research is the gap between how drivers assess their own impairment and how they assess other people’s. Sixty-nine percent of surveyed drivers said they can reliably judge whether they are too impaired to drive after using cannabis. Yet only 43% trusted fellow state residents to make the same call correctly, and just 36% trusted cannabis tourists to do so. The math does not add up — and it rarely does when people assess their own risk versus that of others.
Only 35% of drivers in legalized states said they were confident they knew their state’s cannabis driving laws. One in three did not believe a police officer could reliably detect cannabis impairment behind the wheel. These numbers describe a legal landscape that has expanded faster than public understanding of it, with millions of drivers navigating rules they do not fully know and enforcement tools that are not yet adequate to close the gap.
The self-reported effects on driving paint a clearer picture than the confidence figures. Twenty percent of surveyed drivers recalled experiencing slowed reaction times after using cannabis. Thirteen percent noted greater difficulty concentrating. Eleven percent reported difficulty staying within the speed limit. Eight percent said they struggled to stay in lane. And yet 42% reported noticing no changes to their driving at all — a figure that sits uncomfortably alongside the 47% who said they felt impaired after using cannabis. The two groups overlap in ways that are difficult to reconcile, and that tension is at the heart of what makes cannabis-impaired driving such a persistent enforcement challenge.
A Road Safety Trade-Off With No Simple Answer
The combined picture — less alcohol, more marijuana, fewer impaired driving deaths — might look like a straightforward improvement. But the mechanics behind it are more complex. Alcohol remains far more prevalent than cannabis across all age groups: a 2023 national survey found that 68.1% of adults aged 18 to 25 used alcohol in the previous year, compared to 36.5% who used marijuana. The impaired driving fatality reduction is, at least in part, a function of that gap still being large. If it narrows — and the trajectory of both trends suggests it will — the question of whether cannabis impairs driving to the same degree as alcohol will matter a great deal more than it does now.
The research evidence on that question is still forming. Seventy-three percent of surveyed drivers believe alcohol impairment at the wheel is more dangerous than cannabis impairment. The scientific picture supports the view that high-level alcohol impairment tends to produce more severe driving degradation than cannabis — but it also makes clear that cannabis impairment raises crash risk, that combining the two substances amplifies the danger significantly, and that current recommended waiting periods before driving are inconsistent across states and not well understood by the public.
The shift in American drug culture is real, it is generational, and it is now embedded in the road safety data. Whether the net effect on roads over the next decade is positive depends heavily on whether public understanding of cannabis impairment, and the legal frameworks around it, keep pace with the speed at which its use is growing.
