in that quote from the article, the hypothetical (and eventually, very real) victim of a false positive from less than absolute 100% accuracy is not a drunk.

“If it’s [only] 99.9% accurate, you could have a million false positives,” Carlson said. “Those false positives could be somebody trying to get to the hospital for an emergency.”

and ‘ambulance-worthy’ emergencies are not the only critical and potentially life-saving trips that could be affected by a false-positive. how about the doctor who would treat that person in the ER, or a volunteer firefighter responding to a call, or a parent going to fetch their teen after they call for a ride from an unsafe situation…