When it comes to roadside sobriety tests for THC, currently available blood or urine tests can measure one thing, and one thing only: how much is THC in your system.

What these tests can't do, however, is correlate that measurement to your level of impairment.

Translation: cops can tell if you've smoked or eaten weed at some point during the last few days, but they can't tell whether you're too fucked up to drive.

That's because currently, there is no scientifically proven relationship between blood or urine THC levels and dangerous driving.

As has been demonstrated time and time again by both scientific research and investigative reports, THC is capable of remaining in the system days after smoking or eating it, so there's no way to prove that the THC in someone's blood or piss at the time of their arrest was responsible for their shitty car skills — it could have been there from a party they were at three days ago. Add to that the fact people's individual biochemistries and tolerances for THC vary wildly, and you've got a situation where creating a standard, legal definition of stoned driving is nearly impossible.

But … what about breath THC? Could that be used to determine impairment at roadside stops?

This is the question police have been asking recently as they march on in their quest to develop a device that can accurately measure whether you're too high to cruise around in your step-mom's Jetta.

Hound Labs, an Oakland-based company founded by an ER doctor and deputy reserve sheriff named Mike Lynn, is the latest company to claim they've intervened such a device. They say they've done what no other cannabis breathalyzer company has ever done, which is accurately measure breath levels of THC (there are other cannabis breathalyzer companies like Cannabix, but their technology apparently falls behind Hound Labs').

Hound Labs also claims their special little doohickey is sensitive enough to detect edibles.

And, according to a press release issued by the company, the thing is so good and great that it's the "first to be tested by law enforcement at the roadside" and that it'll "require new legal standards that focus on a breath measurement rather than a blood measurement." After all, it conveys measurements in picograms, whereas current tests measure in nanograms per mililiter of blood.

"There are laws coming," Lynn said. Ooo-eee!

"I've heard these stories before," Dale Gieringer, California state coordinator of the National Organization for the Reform of Marijuana Laws (NORML), told VICE, referencing similar THC breath measurement systems that have cropped up in the past, but failed miserably. 

"They're all the same. None of them [are] going to work," Gieringer added. "THC in the body isn't as clear-cut of a sign of impairment as it is with alcohol."

He called the Hound Labs device "unproven technology."

NORML board member and criminal defense attorney Lenny Frieling backs up this assertion, advising that while your breath, or blood or pee or whatever might test positive for THC during a roadside DUI stop, that evidence isn't always admissible in a court of law.

"While measuring a driver’s breath might lead to a suspicion that ingestion had been recent, it certainly does not provide relevant or useful information about impairment or lack of impairment," he says on his website.

So, if impairment truly can't be proven with breath or blood or any other bodily fluid for that matter, why is law enforcement expending so much energy trying to develop a device to do so?

Well, weirdly, research actually seems to be on their side with the whole breath thing. It's not perfect research with perfect results, but shit — it's enough for companies like Hound Labs to go off of.

Researchers recently published a study in Clinical Chemistry that suggests a breath test may actually be the best way for police to analyze a driver's impairment level because the tests are often only positive for THC immediately after participants got high. A.k.a, the tests show a higher THC concentrations closer to the point of weed ingestion, which could suggest that if a person's breath THC concentration falls into a specific range, they could have recently smoked, and are therefore more likely to be threat on the road.

Police in California recently replicated this same experiment with a little field test of their own using the cannabis breathalyzer: they randomly pulled over drivers as study participants, and found that two who admitted to smoking weed 30 minutes earlier tested positive on the breathalyzer. Other drivers who smoked pot within the previous two to three hours also tested positive. Police didn't arrest any of these people because, again, there's no way to tell if they're actually high, but it did prove that the breathalyzers were fairly accurate.

“We are looking for the least invasive way to obtain information that indicates impairment, which is why we are participating in roadside tests,” said Patrick Walsh, the Chief of Police Lompoc, California where the latter test was conducted. "We don't want to arrest people who are not impaired, and yet we don't want marijuana users driving if they are high from recent use."

And if you're wondering if these results may be skewed by everyday smokers who consistently maintain high levels of bodily THC, you should. People who smoke the most often will be the most likely to be labeled as “impaired” during a roadside test, when in reality, they might not be high at all.

Lynn, for his part, told VICE that, in order to control for results like those, independent researchers will be involved. He said The University of California "will be conducting a rigorous clinical trial, using our device and comparing it to their own gold standard [testing equipment] "— with the intent being "for UCSF to publish this study."

In the meantime, Lynn and the rest of Hound Labs and family will be off gathering data. "We'll have data that other people can use — law enforcement groups and research groups." He thinks that soon, we will all finally know the approximate THC breath levels that mean people are too stoned to drive … despite the fact that actual impairment is, again, currently unprovable.

If you ask us, what you're most likely to see happen with this kind of new technology is that it'll increasingly start to be used alongside the same beloved roadside acrobatics cops subject drunk drivers to. That is to say it might get used on you, just do your best in your stoned haze to remember it's probably inaccurate.

If you want, you can come over and practice your nose-touches and straight-line-walking with us later.