Uber may be able to tell when you’ve had a few too many drinks.
The ridesharing giant has explored identifying drunk passengers with artificial intelligence. According to a patent application filed by the company, Uber is planning to use machine learning to better understand the “state” of a passenger when they request a ride.
The patent—which is pretty vague as to its purpose and full of a whole lot of technobabble describing how the system would theoretically work—lays out a system in which Uber could analyse how people typically interact with the app and determine if they are behaving in an “unusual manner,” whatever that might mean.
The patent said: “The system receives the trip request from a user and generates a prediction about the current state of the user using the computer model.
“To predict user state, the system compares data associated with the trip request to data about past trip requests submitted by the user.
“Past trip information may be parameterized to a profile of the user and identify how the user activity of the current trip request deviates from previous (or “normal”) behavior for that user.”
The patent application went further to explain that the system will help users to learn how to typically use the app, so that it can identify unusual behaviour when a user tap to open the app. The system will work with an algorithm to weigh a variety of factors, including typos from spelling when a users is typing his or her address, how precisely a user clicks on links and buttons, walking speed, and how long it takes to request a ride. The time of day, and where a ride is requested may also be considered which are the key factors when detecting an unusual behaviour from users. The patent also added that drivers may be warned of their passenger’s state beforehand to avoid misinformation.
For example, Someone who hails a car to get home after hitting the bars probably won’t be the best at typing out their home address or walking too quickly (if at all). Factors like these are plugged into Uber’s planned algorithm, along with details about when and where the ride was requested. A late-night Uber call or one that’s from an area loaded with bars, would tip the system off about a rider’s sobriety differently than a daytime call from the suburbs.
In accordance to a patent software noticed by CNN, Uber has drawn up a plan for AI that detects a person’s drunken conduct to accommodate them earlier than they get into the backseat. The system, described by members of Uber’s Belief & Security workforce in 2016 and revealed yesterday, tracks how somebody sometimes makes use of Uber’s app: how rapidly they kind (and with what number of typos), how exactly they click on on the buttons, their strolling pace, and the way in which their telephone is usually held or dropped on any given day. Uber will then either warn drivers, match riders with drivers who have relevant training or not under the influence of alcohol or deny the request for a another ride.
The new AI system when implemented will help reduce the high rate of violence on Uber’s driver, because over the past years there have been over 100 cases of abuse and assault by Uber drivers in the past four years – and in many of these cases, passengers were drunk, according to patent report filed.