25.8 C
New York
Wednesday, May 22, 2024

Driving While Baked? Inside the High-Tech Quest to Find Out

Everyone knows about 4/20, but true stoners also observe 7/10. In the spirit of a ninth grader entering 58008 on a calculator and flipping it over to spell out “BOOBS,” the date 7/10 looks like the word “OIL” upside down and backward. This makes the 10th day of July an ideal occasion to celebrate the popular forms of concentrated cannabis oil known as dabs or wax, which can be three or four times as potent as regular bud.

In 2017, Doug Fraser spent six months planning a 7/10 event at the legal weed store he manages north of Seattle. He booked food trucks and glass blowers, and more than a thousand people showed up. Fraser worked a 12-hour shift that day, though he did not get stoned, as state regulations do not allow cannabis industry employees to partake on the premises. Before he went to bed that night, he took a small hit of oil—a modest reward for a hard day's work.

The next morning, Fraser woke up early and drove back to the pot shop. When he clocked out at 6 pm, he couldn't wait to get home. His 1998 Volvo was barreling up Interstate 5 at 81 mph when a cop pulled him over. Fraser felt silly for rushing, and when he saw the police officer eyeing the employee badge hanging around his neck, he felt worse than silly. The badge featured his store's logo, ubiquitous on local billboards.

“When was the last time you smoked?” the cop asked.

“Not today,” Fraser responded, his anxiety spiking.

The officer asked him to step out and complete a field sobriety test on the side of the highway. Fraser performed six tasks, including counting out a duration of 30 seconds in his head, as if to test his inner stopwatch. When the cop asked him to balance on one leg, Fraser complied, but explained that an accident had left him with fake knees, bolts in both femurs, and restless leg syndrome. In the face of the cop's skepticism as he completed test after test, Fraser grew so nervous he started to shake. This seemed to confirm the officer's suspicions. Fraser was arrested for driving under the influence of marijuana. The cop squeezed his wrists tight as he secured the handcuffs.

From the highway, Fraser was taken to a hospital to have his blood drawn. According to Washington law, it is illegal to drive with 5 nanograms or more of THC—the compound in pot that causes a high—per milliliter of your blood. That might sound like a straightforward, objective test, but unlike alcohol, which is water-soluble and dissipates as you get less drunk, THC is fat-soluble, meaning it might fade within hours but may also linger in the body for weeks after the high has passed—longer than any other recreational drug. Although Fraser had last consumed marijuana about 20 hours earlier, his blood contained around 9 nanograms of THC per milliliter.

“I was almost twice the legal limit,” Fraser recalls. “I was blown away.” Like most people, he had no idea it was possible to be arrested for driving under the influence almost a full day after getting high. That evening, Fraser discovered the hard way that the science behind stoned-driving laws makes absolutely no sense.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

There is no such thing as a weed Breathalyzer, no biological factor that indicates whether someone is impaired by cannabis at the moment you test them—not blood or urine or hair or breath or spit. While every state prohibits driving under the influence of weed, no state has found a reliable way to sort the stoned from the sober. Like Washington, some states have implemented the 5-nanogram-per-milliliter threshold, but cannabis pharmacokinetics are so variable that even if two people share a joint, one person might dip below that level within two hours and the other might stay above it for the rest of the week. Stoned driving is therefore one of the biggest unresolved sticking points in the long slide toward legalizing marijuana in the US—a Kafkaesque quandary with no clear solution.

Driving high is not as dangerous as driving drunk, but marijuana can interfere with essential skills, including reaction time, processing speed, and visuospatial perception. Measuring marijuana's overall impact on road safety, though, is statistically fraught. In 2020 a study found that in the years after Washington state legalized marijuana, the percentage of people involved in fatal crashes who tested positive for THC doubled. The study made national headlines and was cited by politicians who raised fears about stoned driving as a reason to keep weed illegal. But while those findings probably do reflect the fact that more people consume cannabis when it's legal, they didn't show that THC impairment caused the crashes, merely that the substance was present. Washingtonians who tested positive might have gotten stoned days before they put their key in the ignition. (The total number of road fatalities declined.)

Authorities sometimes acknowledge that marijuana toxicology reports are junk, but that often means defaulting to another flawed method of assessment: human judgment. A third of stoned-driving arrests in Colorado rely only on the opinion of a police officer—just a cop on the side of the road, considering who to arrest and who to let go.

This means a whole lot of people are going to jail, paying fines, and possibly losing their jobs, their driver's licenses, and their children for driving stoned, even though they were driving sober. The exact number is difficult to quantify because most places do not track DUI arrests by type of intoxicant, but in the 18 states that have legalized recreational pot for adults, this failure of science and regulation amounts to a continuation of the war on drugs.

Doug Fraser ended up with a speeding ticket, 15 days of house arrest, three years of probation, and around $4,000 in fines and fees—a typical sentence. Fraser is white, but across the country, drivers with darker skin are disproportionately getting busted for driving around with THC in their veins. People of all races use pot at similar rates, but Black people are about four times more likely to get arrested for possession. Black people are also significantly more likely to get pulled over and more likely to be searched once they're on the side of the road, but less likely to have drugs in the car.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

As cannabis legalization spreads, state governments often look to Washington and Colorado—the first places to allow recreational marijuana—for guidance on how to write the rules. Both states passed the 5-nanogram THC limit as part of legalization, but Fraser is hoping he can keep that threshold from serving as a national model. In an attempt to overturn Washington's standard, and maybe even stop it from becoming the law elsewhere, Fraser's lawyers argued his case before the state supreme court on January 18, 2022. A ruling is expected this summer.

But what should take the place of the blood toxicology standard if THC presence doesn't always signify impairment? “A scientifically backed field sobriety test specific to cannabis—that's what I would like to see,” Fraser says. “There are always going to be rule-breakers, and there definitely needs to be a test for, ‘Hey, are you impaired right this second?’ Whether you smoked 20 minutes, 20 hours, or 20 days ago, that test should exist.”

Fraser is not the only one looking for such a test. Pretty much everyone agrees that the 5-nanogram threshold isn't working: cops, prosecutors, toxicologists, social justice activists, potheads. To replace it, though, someone has to come up with an easy, portable way to attain something that has more or less eluded science for millennia—a good, clear glimpse into a human's state of mind.

About 15 years ago, Charles Duffy, then a neurobiology researcher at the University of Rochester Medical Center, set out to build a device that could quickly and effectively assess cognitive impairment. Because CT scans and MRIs cannot yet detect definitive biomarkers for neurodegenerative diseases like Alzheimer's, cognitive evaluations are usually performed by doctors and are therefore subject to error and bias. Duffy's idea was to create a technology that would eliminate the element of human fallibility, tracking the cognitive decline associated with aging “independent of the patient's language capabilities and cultural, ethnic, educational, or socioeconomic background.” The device would evaluate memory, visual acuity, processing speed, reaction time, and higher-order reasoning and concentration skills known as “executive function.” He wanted the 10-minute test to shift in difficulty and speed according to the patient's abilities, using a method of statistical analysis called adaptive psychophysics. Duffy called his device Cognivue. An early model weighed 145 pounds and resembled a microfiche machine, with a monitor and joystick in a five-sided metal box, atop a 4-foot-tall rolling cart.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

When the FDA cleared Cognivue for use in 2015, the technology was so unusual that the agency had to create a new category of medical devices, called cognitive assessment aids, to classify it. Duffy hoped to get the Cognivue device into the offices of neurologists and primary care physicians, but found he was better at research than sales.

Three years later, he sold the company to a billionaire HR software tycoon named Tom Golisano. Golisano brought in a new team that helped slim down the device. It now resembles an Apple iBook circa 1999, weighing 8 pounds and opening to reveal a three-sided glare hood that folds out. Instead of a keyboard, there's an inset wheel, controlled by a small, rotating knob on its outer perimeter, kind of like the mechanism for operating a salad spinner.

This wheel and knob are crucial to Cognivue's algorithm, which moves so quickly it can't run on Wi-Fi or a typical computer. In response to questions like “Which shape is different?” and “Which word is real?” the subject maneuvers the knob to move a green circle on the screen until it's aligned over the correct answer. Take too long to respond and the test moves on to the next prompt. As the assessment progresses, the sensitivity of the wheel changes—now rotating as if through peanut butter, now flying like a bike in first gear—evaluating the subject's ability to adjust.

To measure how well a person can see and respond to visual stimuli, Cognivue makes trippy use of flashing, frantic dots. “Stay on the bright dots,” the test commands early on, offering a green wedge to hover over the white twinkling specks migrating around a circle of gray twinkling specks. Easy enough, until the white specks dim and threaten to disappear. Later, to evaluate the subject's ability to distinguish motion, six circles of dots appear, each a sparkling universe of chaos, electrons bouncing against the walls. Some dots bubble upward, others cycle clockwise, and the test asks, “Which dots move differently?”

Most of the tasks begin as relatively easy for any functional adult, but as with any measure of the brain's abilities, there is a wide range of “normal” scores. Cognivue has screened thousands of healthy individuals to determine how many mistakes the average person might make when taking the test, or how many milliseconds of delay might imply something is wrong.

Bolstered by Golisano's billions, Cognivue's sales reps began trying to place the device with doctors who might see patients in the early stages of cognitive impairment: neuropsychologists, audiologists, general practitioners. The reps also considered how the company might boost its bottom line by attracting the general public. Would people be tempted to take a Cognivue test while waiting to fill a prescription at a pharmacy? What about at a corporate wellness fair? Experimenting with branding the device in different environments, the company found that the average person was not interested in a “cognitive test” but was eager to sign up for a “brain health screening.”

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

Cognivue also soon realized that the device had the potential to be an especially robust form of what experts call an external impairment test for cannabis: something that determines whether someone is high by evaluating their behavior and abilities, rather than measuring THC in their body. A field sobriety test is essentially a series of external impairment tests. The difference with Cognivue is that the entire test is both quantitative and automated, so the results are not filtered through the perceptions of an individual's judgment.

Considering the racial bias in how cannabis laws have been enforced, activists tend to argue that the gold standard for an impairment test would be one that does not rely on the subjective opinion of a police officer. The National Organization for the Reform of Marijuana Laws (Norml) released an app in 2015 called Canary that was meant to provide an example of how a cannabis impairment test might work. Like the field sobriety test that Fraser took on the side of the highway, it involved counting and then standing on one leg. Weed causes people to perceive time as passing more slowly, so the counting test is a popular indicator of cannabis consumption, though of course not every sober person has a superb internal metronome. The Druid app, an impairment test developed by a psychologist at the University of Massachusetts Boston, also asks people to balance and to estimate when 30 seconds have passed.

Cognivue's device involves significantly more advanced technology. But does it truly work any better than blood tests or counting or standing on one foot? To figure that out, the company needed a clinical trial. And to design one with a real chance of being taken seriously, it needed to bring together the two camps that care most about the details of how weed DUIs are measured and prosecuted: law enforcement and stoners.

Cops, the more intransigent stakeholders, came first. In 2019, an internist who once led a Doctors Against Distracted Driving campaign began using the Cognivue device with his elderly patients in Rochester, New York. After a while, he mentioned it to some of his law enforcement and political contacts as a potential tool for evaluating impaired drivers. At the time, the state was beginning to consider how to legalize cannabis. Word of the test's relatively sophisticated assessment capabilities spread from sheriff to politician to bureaucrat. Soon, the assistant commissioner for the Governor's Traffic Safety Committee, Chuck DeWeese, was on the phone with Cognivue. An employee recalled DeWeese saying that “if this works, your owner is going to be a billionaire.”

Any change to how cannabis impairment might be evaluated in New York needs to go through Renée Borden. Borden is the state coordinator for New York's drug recognition experts, elite cops known as DREs. Most impaired driving arrests happen after brief roadside evaluations; if alcohol is ruled out, the driver may get a more thorough examination by a DRE, who has training in how to recognize the effects of seven classes of intoxicants. How accurate these evaluations are for cannabis impairment is unclear; The Boston Globe recently described DRE protocols as “little more than pseudoscience.”

Borden is the kind of law enforcement officer who tells me she is worried teenagers are soaking scrunchies in drugs to inhale during class. She has the hardened air of a woman who has excelled in a profession dominated by men, with an immaculate mani-pedi and tasteful silver jewelry but otherwise casual in jean shorts, flip-flops, ponytail. She seems confident in how her specialized police officers are already assessing marijuana impairment, even though she recognizes there is no scientific way to completely verify their judgment. “We've talked to doctors who said, ‘If you really want to know if they're impaired, you need a slice of their brain,’ so that's not worth it,” Borden tells me. “DREs are our technology.”

In late 2019, as plans for a clinical trial of Cognivue began, an instructor who worked for Borden—and was following the stoned-driving issue closely—told folks from Cognivue that if they wanted to do this trial right, they needed to reach out to a scientist in Colorado named Frank Conrad. Conrad operates at the fulcrum of legal weed and legal weed science, two worlds that rarely interact, mostly because of federal law.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

In order to do legitimate, peer-reviewed research involving cannabis in the United States, scientists have long been required to use the notoriously weak and moldy government weed grown at the University of Mississippi. Gaining permission to do clinical trials involving pot and humans often involves years of bureaucratic hurdles. The studies that do happen tend to involve mice, not people. This is a part of why we have so little data about marijuana's medical potential, or even about how THC interacts with common prescription medications. At the same time, cannabis is legally available to consumers and patients in a majority of states, enabling scrappy or profit-driven explorations of the drug's potential: companies conducting private trials for competitive advantage, say, or off-grid pot farmers supplying cancer patients with THC goo.

In the middle, balancing reliable data with the realities of the Wild West cannabis market, are people like Conrad, a bald, 43-year-old molecular biologist and stoner with a big, broad grin, a guy who drives a burnt-orange Honda hatchback and wears hiking boots to meetings where everyone else is in business casual. Conrad started an early weed lab in Colorado that did groundbreaking research on the dangers of cannabis pesticides, prompting a change in how states write their regulations. He often works as an expert witness to help defend people who test positive for THC after using nonintoxicating CBD products, leaving him infuriated by all the ways faulty science and confusing laws can trap people in a criminal justice nightmare. When Conrad heard what Cognivue was up to, he immediately agreed to help. “I've been waiting for you guys for six years,” he said.

Conrad is Borden's foil. He questions cop methodology; she teaches it. He mistrusts any government science based on the Mississippi-grown schwag; she relies on it. He believes too many people are being arrested for stoned driving; she believes it's too few. He wants Cognivue's trial to succeed, and to reduce human subjectivity in stoned-driving arrests; she doubts some newfangled technology can improve the situation.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

In order to get state approval for its device, Cognivue needs the support of both law enforcement bigwigs like Renée Borden and cannabis advocates like Frank Conrad. The company's success hinges on a delicate détente between two cultures that traditionally despise each other. “It's like prosecutors and defense attorneys: They have different goals,” says Kristin Weber, Cognivue's director of safety channels. “We walk the line in the middle.”

If Cognivue could show that scores on their computerized test demonstrated impairment better than the highly trained cops or THC in the blood, New York just might implement the first objective measure for stoned driving in the United States, potentially saving many thousands of people from unnecessary arrest.

“This is a weird paradigm shift,” says Conrad. “If we survive these initial tests—and I'm confident we will—this will gain critical mass in the next two years across the country.”

By the time New York state legalized pot in March 2021, plans were already set for a Cognivue study in Denver over a weekend in early June. Because the challenge of measuring cannabis impairment is greatest among daily cannabis users, who nearly always have some THC buildup in their blood, all 48 people recruited for the trial were potheads. Many heard about the study on Reddit; the post offered $150 and a free vape pen battery in exchange for testing “cognitive performance before and after THC inhalation.” The plan was to evaluate the subjects when they first arrived—using a blood test, a Cognivue test, and a New York state DRE test administered by law enforcement—to establish each person's baseline. After a lunch break, the trial participants would all get stoned and repeat the same assessments, over and over, until they were back to feeling how they did when they arrived.

To get around the requirement that human cannabis studies use the crappy weed from Mississippi, Conrad decided trial participants would bring their own vape cartridges, figuring that THC measurements from repeated finger-prick blood tests would provide a better and more easily quantified control variable than everyone getting high on the same product. (Because of variations in how we inhale, two people could hit the same pen the same number of times and have differing amounts of THC in their systems.)

I show up to the trial's location, a private cannabis consumption lounge in eastern Denver with plenty of outdoor space, psychedelic art, and encouraging signs with messages like “DON'T WORRY IT'S LEGAL BE HAPPY.” The 10 cops that Borden brought from New York for the clinical trial are white men, most with hulking biceps and a self-serious demeanor. The gathered stoners are scrawnier, scruffier, and more diverse. They are predominantly male but closely reflect Denver's demographic makeup, which is mostly white and light-skinned Latino, with some Black and East and South Asian folks. (Because of ethical research restrictions, I agreed to not share the names of trial participants. The drug recognition experts asked me to use first names only.)

Scattered in patio chairs around the lounge's graveled courtyard, the DREs spend their breaks chatting about jiujitsu and people who huff the compressed gas product Dust-Off. The potheads discuss air fryers and cookies. Without the usual context of a possible arrest, the vibe between the cops helping out with the trial and the stoners who signed up as subjects veers from surreal to downright chummy, as if Tom and Jerry took a break from the endless chase to discuss the finer points of mousetrap methodology.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

While her underlings socialize, Borden stays glued to her laptop, reviewing grant applications and then watching TV. Conrad, on the other hand, bounces from group to group, ecstatic, regaling subjects with tales of courtroom drama and forensic weed science. After lunch, when it's time for the subjects to get high, Conrad gathers the participants together, telling them to vape however long they need to be nice and stoned.

“Proceed!” he shouts, and the only sound in the courtyard is the hiss of inhalation, followed by a chorus of coughs. Over the next hour, the clinical trial gets wacky. Eyes turn red, phones get lost, giggles abound. Conrad has the subjects place their vape pens in a sealed envelope until the trial finishes. A guy with shaggy hair tries to lick his shut, then realizes it's the kind with a peel-off label.

Once stoned, the study participants find the 10-minute Cognivue test overwhelming, to say the least. “I kept questioning my sanity,” one guy tells me. The clusters of vibrating dots confuse and frustrate almost everyone. “Are there dots? There are not dots,” one person says. “All the dots, they turned into an amorphous borb,” adds another. A guy with long hair sums up the experience: “I hate them dots.”

Afterward, the DRE assessments take about 45 minutes and include everything from a pulse check to an ultraviolet scan of the tongue. One man, asked to count out 10 paces, begins at “three,” realizes what he's done at “five,” then breaks into laughter. A cop explains to a subject that one thing he's looking for is “muscle charmers”—a fluttering spasm, often in the eyelids or neck, that indicates recent cannabis consumption. “I imagine seeing it when you're high is pretty cool,” he says.

The next round of testing, just over an hour after the vape session, is more contentious. The subjects are convinced their high has faded, but the DREs expect that everyone is still intoxicated. The National Highway Traffic Safety Administration asserts that cannabis impairment peaks between 90 and 120 minutes after smoking or vaping. From the cop perspective, this is one of the greatest challenges to catching stoned drivers. “After 90 minutes, impairment goes up,” a DRE named Guy tells me.

Conrad thinks this is bullshit and based on “crazy research” using low-THC Mississippi joints that take 30 minutes to get someone high, which is not how most people become stoned in the real world, puffing slowly on weak pot. The Mississippi joints sit in a freezer for years and include ground cannabis stems and stalks—a move the commercial marijuana industry would consider akin to mixing strips of orange peel into a fruit salad.

Nearly two hours after the subjects vape, Guy is deep in a convivial discussion with the woman he's evaluating over the need to approach addiction with more treatment and less punishment. Then he shines a light in her eyes and says, “You got big pupils.”

“I am sober,” the woman responds, a sudden edge in her voice.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

“You're not sober,” Guy says.

“I'm as sober as you'll get me in the next week or two.”

“Well,” Guy laughs, “it's all relative.”

For many of the subjects, their motivation to participate in the trial is about more than just saving sober people from unnecessary arrest. It's about proving that experienced potheads are safe on the roads and should only be arrested if an external measure like the Cognivue test shows they are impaired, regardless of the last time they smoked. To put it plainly, a lot of the trial participants say they drive high.

“I purposefully get high to drive on mountain highways,” a woman in a Deadpool jersey says. “So beautiful and scenic!”

“It makes me drive more carefully, to be honest,” says another woman. “I drive slower.”

This is a common belief among daily marijuana users—that driving stoned is just as safe as driving sober, if not safer. Research does show that experienced users often overcompensate for perceived impairment with increased caution and concentration. Accordingly, several study subjects think they did best on the Cognivue test taken right after they vaped—as they were “super zoned in,” as one woman described it.

“I'm looking forward to them seeing people's scores improve,” a guy in a kilt tells me. “Take that data to Congress!”

The drug recognition experts find this confidence alarming. Research from the National Institute of Drug Abuse suggests that even after the high ends, as long as there is THC in someone's system, the person is twice as likely to get into an accident. For daily cannabis users, that would mean that even if they were to stop smoking, their psychomotor ability could be impaired for up to three weeks.

“Even though you feel normal, there's physiological things you can't control,” a DRE named John tells one study subject.

“But I've been driving for many years,” the subject counters, saying he knows how to handle the roads while stoned. His eyes are bloodshot, but he insists he would be fine to get behind the wheel. John asks him to extend his arms and then touch his nose, one hand at a time. The subject misses his nose. John asks him to walk in a straight line, and the guy wobbles, saying, “Whoa, a little shaky on that one!”

Later, after the subject is out of earshot, John tells me he doesn't even know where to start with people like this. “After decades of education, everyone knows driving drunk is dangerous, but 40 percent of smokers say it doesn't affect their driving,” he says. “How do you fight against half the people?”

Many of the study participants, likewise, think driving stoned in Colorado is unlikely to cause either an accident or an arrest. “The cops are so nice here!” multiple people tell me. “They're chill with weed.” A white girl with short red hair says she once got pulled over while smoking a blunt, and the cops let her go. A white guy in clear-framed glasses says he once got pulled over immediately after hotboxing the car, and though the police officer gave him a ticket for speeding, his only response to the pungent stink of pot was to say that next time he should wait four hours before driving. A friend of his, the guy goes on, once fell asleep at a stop sign with a bong in the passenger seat; when he woke up, a cop was gently suggesting he walk home. Many of the white people, Asian people, and light-skinned Latino people at the clinical trial have stories like this.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

Hearing this, the handful of Black subjects roll their eyes. A Black man in an Ed Hardy shirt tells me he never smokes in his car because he gets pulled over two or three times a month—more if he's driving his mother's BMW. He keeps his dash cam on at all times; he once had to use the footage in court to undermine the testimony of a cop who gave him a field sobriety test and then lied in court about how it went. Another Black man tells me how earlier this year he was in the passenger seat when his friend, also Black, got arrested for driving under the influence of cannabis, though the friend had been completely sober.

John acknowledges that, for all their specialized training, DREs are still subjectively evaluating people. “Everyone has a different personality. The ability to put people at ease is different,” he tells me. But that doesn't mean he thinks that the Cognivue test or any other test would be better: “There is no scientific method, period.”

This attitude frustrates Conrad. “What you really want is not to have a human doing this,” he says. The DREs disagree.

“I'm not afraid of somebody replacing what we do,” John says, “because they can't.”

“Replace?” Guy says. “Ain't no fuckin' chance.”

“That would be a scary day,” intones Rick, another DRE. “If you lose the human element.”

“Big Brother!” Guy adds.

At the same time, the drug recognition experts often trash-talk the skills of average cops who make the majority of stoned driving arrests. They emphasize how much more DREs know, how, as Borden tells me, they're “the only ones with the title of expert.” About 350 of New York's 63,000 law enforcement officers are trained as DREs, and around 2,400 are trained in a hybrid program that involves the standard field sobriety test. Even if the DRE methodology were flawless, it would be impossible for the current roster to handle the state's full load of stoned-driving evaluations. And with strict standards in place for who can become a drug recognition expert, we're unlikely to see significantly more DREs on the streets any time soon.

When the clinical trial is over and a group of stoner subjects is waiting for their rideshares, I ask them which method they thought was better at figuring out they were stoned: the cop or the Cognivue test? Nearly everyone said the DREs were better. “But I think the cop was looking for something to be wrong,” adds the only Black person standing there, a woman. “The computer is at least unbiased.”

When the results of the Cognivue clinical trial come back a few months later, it turns out the DREs were at times relatively lenient in their evaluations. During the initial baseline testing, before the vaping session, the DREs marked the fewest people down as impaired. According to the blood test, 74 percent of the potheads were stoned at the start of the day; according to the Cognivue test, 47 percent were impaired; and according to the DREs, only 21 percent were stoned.

Most PopularBusinessThe End of Airbnb in New York

Amanda Hoover

BusinessThis Is the True Scale of New York’s Airbnb Apocalypse

Amanda Hoover

CultureStarfield Will Be the Meme Game for Decades to Come

Will Bedingfield

GearThe 15 Best Electric Bikes for Every Kind of Ride

Adrienne So

But an hour after the subjects vaped, when many said they felt sober again, the Cognivue test found that about half of the potheads were indeed no longer impaired. According to the blood tests, 84 percent were still above 5 nanograms of THC, and according to the cops, 68 percent were still impaired.

The aspect of the Cognivue test that most clearly indicated subjects were impaired was those roving, fading dots. The errant specks measured “visual salience reaction time,” meaning that getting high slowed down the potheads' ability to understand what they were seeing, a crucial skill in driving. But in the tests taken immediately after getting high, a few outlier participants did in fact improve their reaction time over the baseline, showing again that weed affects everyone differently. (One person went up to 2,091 nanograms of THC per milliliter of blood after vaping; another peaked at 150 nanograms and was back down to zero within an hour.) So even if the Cognivue test is the best available assessment when it comes to determining who is impaired and who is not, the trial made it clear that within a group of potheads, inconsistency is the only constant. Some people can handle driving while high, and some can't.

Cognivue is now in talks with the Colorado Department of Transportation to potentially help the state in its efforts to police stoned driving, and the company is planning future clinical trials, including one that will involve comparing how subjects fare on its test after consuming alcohol, so the state can see how the device measures impairment from a more familiar intoxicant. Considering how slowly governments tend to move, it could take years before Cognivue is actually used by law enforcement.

Yet even if Cognivue's test were to be implemented nationwide, that wouldn't solve a crucial issue the trial made clear. Many daily cannabis users are unaware of their own impairment. This is why people like Doug Fraser, whose weed DUI case is being decided by the Washington Supreme Court, want cops to use an impairment test like Cognivue's instead of a blood THC measurement like the one that got him convicted. They believe that if they are sober, they will pass. But almost half of the potheads in this study had a delayed reaction time before they even vaped. And although a few people told me they got high that morning, before coming to the trial, most said they did not, so these baseline results may imply that many stoners have perpetually slow reaction times—a possibility that aligns with the National Institute on Drug Abuse data showing that any amount of residual THC in someone's system makes them twice as likely to get into an accident, regardless of whether they feel stoned or sober.

If Fraser's case gets the 5-nanogram limit overturned in Washington, it will surely save many people from arrest. And perhaps Washington state will follow New York and Colorado by taking a closer look at the Cognivue test as a more reasonable alternative.

Ultimately, the problem for people who like to smoke pot every day is that they have no real idea what kind of stoner they are. Do you think you're fine to drive high but are, in fact, too slow to react even when sober? Or are you actually fine to drive a mere five minutes after hitting the bong?

The stories we tell ourselves about who we are and what we're capable of can shape our lives in profound ways. But when it comes to people who regularly use cannabis, no matter what assessment method becomes standard in the coming years, the choice to get behind the wheel could always be a risky one.


This article appears in the March, 2022 issue. Subscribe now.

Let us know what you think about this article. Submit a letter to the editor at mail@wired.com.


More Great WIRED Stories📩 The latest on tech, science, and more: Get our newsletters!They were “calling to help.” Then they stole thousandsA new trick lets AI see in 3DLooks like folding phones are here to stayWomen in tech have been pulling a “second shift”Can super-fast battery charging fix the electric car?👁️ Explore AI like never before with our new database💻 Upgrade your work game with our Gear team’s favorite laptops, keyboards, typing alternatives, and noise-canceling headphones

Related Articles

Latest Articles