Somewhere in this country, almost every day, the same scene repeats itself. A driver, allegedly over the legal limit, puts a vehicle into something it was never supposed to hit — a parked car, a mailbox, someone’s driveway. Field sobriety tests get declined. The breath test gets refused. Handcuffs go on. Local news covers it for a day, then moves on to the next one.
What almost none of those stories mention is that Washington already tried to legislate this exact scenario out of existence. The law passed in November 2021. It came with a hard deadline. That deadline is now well over a year in the rearview mirror, and the reason your car still can’t tell whether you’re drunk has nothing to do with politics stalling in some committee. It has to do with a definition problem buried inside the statute itself, one that eliminated the simplest fix before engineers ever got a real shot at building it.
The Law Nobody Noticed
The Auto Wire covered the early version of this story back in 2024, when it looked like lawmakers simply wanted a breathalyzer built into new cars. The truth buried in the statute is stranger, and far more interesting.
The law doesn’t order a breathalyzer. It orders the National Highway Traffic Safety Administration to publish a new Federal Motor Vehicle Safety Standard requiring advanced drunk and impaired driving prevention technology in every new passenger vehicle sold in the country. NHTSA’s deadline to finalize that standard was November 15, 2024.
That deadline has already passed.
What NHTSA actually published by that date was an advance notice of proposed rulemaking, federal-speak for “we are still asking questions before we even draft a rule.” The document runs more than two hundred pages and mostly consists of NHTSA asking the public and the auto industry to help it figure out what the law even wants built. Two years after passage, the agency’s own paperwork admits the mandated technology doesn’t fully exist yet, in production form, for a private car.
Washington Banned The Easy Answer
Here’s the detail that makes this genuinely strange. Congress didn’t just require alcohol-detection technology, it required that the technology work passively, defined in the law as functioning “without direct action from vehicle occupants.” A driver blowing into a tube counts as direct action. So does pressing a button to submit a sample. By writing “passive” into the statute, Congress effectively disqualified the one alcohol-detection technology that already exists, is already certified, and is already sitting in courthouse parking lots strapped into the steering columns of repeat DUI offenders across the country: the directed-breath ignition interlock.
The rule doesn’t ban ignition interlocks. Those remain standard punishment for DUI convictions in dozens of states. What it does is exclude that approach as the technology every new car must have. A driver who has never been arrested, doing nothing wrong, would still have to interact with a directed-breath device every single time the car started. NHTSA has openly worried that would annoy sober drivers into disabling the whole system, which under the law’s own safety standard would make the requirement legally impracticable in the first place.
What Engineers Are Actually Trying To Build Instead
Since 2008, NHTSA has co-funded a program called the Driver Alcohol Detection System for Safety, or DADSS, run jointly with the Automotive Coalition for Traffic Safety, a group representing essentially every major automaker. DADSS is chasing two very different pieces of hardware, and both sound more like something out of a hospital than a parts catalog.
The touch system shines near-infrared light into the skin of a driver’s palm through a sensor built into a push-button starter, steering wheel, or gearshift, then calculates blood alcohol concentration from what bounces back. The breath system tries to read alcohol and carbon dioxide out of the ambient air around a driver’s face without requiring so much as a puff, a much harder engineering problem than the directed-breath interlocks already on the market, because it has to isolate one person’s exhaled breath inside a moving cabin full of air currents, vents, and other passengers.
None of it is close to sitting in a showroom. A directed-breath version, built for zero-tolerance commercial fleets rather than the general public, was licensed out in December 2021 and is already logging miles with trucking fleets. A fleet-grade touch system was targeted for 2024. Reference designs actually rated for the legal limit and suitable for a private passenger car, the version that would eventually go into a Camry or an F-150, weren’t expected before 2025 according to the program’s own projections, and NHTSA has publicly noted those dates could slip further. None of that timeline includes the years automakers still need to design the sensors into an actual production vehicle.
We’ve Basically Run This Experiment Before
This isn’t the first time Washington tried to build a car that refuses to cooperate with a driver’s bad behavior. In 1974, federal rules required new cars to have an ignition interlock tied to the seatbelt: buckle up, or the engine simply wouldn’t start. Drivers hated it. The backlash was loud enough, and fast enough, that Congress banned the government from ever mandating that kind of interlock again, within about a year of the original rule taking effect.
NHTSA’s own rulemaking document seems to know exactly how this story ends if it isn’t handled carefully. It cites federal appeals court precedent warning that a safety system doesn’t count as practicable if drivers simply find a way to turn it off. That’s not a hypothetical concern. It’s the agency’s own history repeating, fifty years later, with alcohol sensors instead of seatbelt buzzers.
The Law Has A Loophole Built Into It
Federal law bars automakers, dealers, rental companies, and repair shops from disabling a required safety system once it’s installed. It does not bar individual owners from doing exactly that. NHTSA’s own document acknowledges the gap directly: nothing in the underlying Safety Act stops a private owner from removing or defeating the very system Congress is requiring automakers to spend billions building. State inspection and tampering laws may complicate that in practice, but the federal loophole is real, and it sets up the same kind of gray-market defeat-device culture that already exists around diesel emissions equipment, except this time the thing being deleted would be a sensor that reads blood alcohol.
Nobody Has Solved The False-Positive Problem
Federally funded research out of the University of Iowa, testing a similar impairment-detection approach, produced an uncomfortable result: in some cases, drowsy drivers who had consumed zero alcohol got flagged by the algorithm as alcohol-impaired. Stack that against the legal requirement that any mandated system be practicable, and you get a genuine engineering and liability problem. A system that occasionally strands a stone-sober, simply tired commuter creates its own safety risk and its own lawsuit. A system tuned to avoid that ends up letting more actual drunk drivers through. NHTSA has not said which failure mode it’s willing to accept, because there isn’t a comfortable answer.
That tension connects to a bigger pattern this publication has been tracking for a while: cars are increasingly being asked to police their own drivers, not just protect them. We’ve written about how driver-monitoring cameras could eventually help decide whether someone is fit to drive at all, and about how difficult it already is to keep a driver with a suspended license off the road even with today’s much cruder enforcement tools. Virginia has already moved down this road at the state level, letting judges order court-mandated speed limiters instead of an outright license suspension. A biometric sensor built into the steering wheel is a much bigger step than either of those, and it’s arriving on a timeline nobody fully controls.
Why This Should Matter To You, Not Just Repeat Offenders
It’s worth remembering what a refused breath test actually does today, right now, without any of this new hardware. Every state operates under some form of implied consent law: get a license, and you’ve already agreed that police can test you for impairment, with real consequences, usually an automatic license suspension, if you refuse. A passive, always-on sensor built into the car itself doesn’t ask for consent at all, because it never asks a question in the first place. It just measures. Whatever legal comfort exists today in declining a test evaporates once the vehicle itself becomes the instrument doing the measuring, continuously, as a condition of starting the engine.
That’s not a small shift. It’s a fundamental change in the relationship between a driver and a machine that, for more than a century, has only ever needed a key and a pulse to run.
What To Remember When This Comes Up Again
The federal deadline for cars that can stop drunk driving before it starts has already come and gone. The most recent public action on the file, a request to approve paperwork for testing what the in-car warning light should even look like, landed in the middle of 2026, nearly two years after the law said the whole system should already exist.
Detecting a drunk driver was never really the hard part. Breathalyzers have done that reliably for decades. Building a sensor accurate enough to trust, invisible enough that sober drivers tolerate it, and legal enough to survive the same backlash that killed the seatbelt interlock fifty years ago, that’s the actual assignment Congress handed the auto industry. Nobody has finished it yet. Don’t expect your next new car to finish it either.