Will Uber use artificial intelligence to identify drunk passengers?

Wednesday, 13 Jun, 2018

It can identify "a normal or abnormal state of the respective user of the respective service", the patent reads, based off training a computer model to look at certain data sets, such as the angle the device is held at, the time the vehicle is ordered, or walking speed, according to the patent, entitled "predicting user state using machine learning".

However, it is not exactly clear if this data will be used to deter an Uber driver from picking up a passenger in order to evade any potential drunken predicament.

It also looks at how many times they tap on the Uber app, exactly where they tap, their walking speed and how long it takes for them to request a ride. The time of day the ride is requested and pickup location will also be factors in the final outcome.

Uber has submitted an application to patent an artificial intelligence technology that identifies drunk passengers. Once the algorithm is created for a user, the app will be able to detect unusual behaviour.

The ride-sharing company said in a recent patent application that it was developed a way to tell "the state of the user". A majority of the cases involved Uber drivers allegedly exploiting drunk young women who were trying to get home after a night out.

For Uber drivers, dealing with drunk passengers is a drawback of the job.

By knowing this information, a driver may be warned of their passenger's state - even before they pick them up. Furthermore, it could prevent the user from joining a trip with other riders, for example with Uber Pool.

Theoretically, knowing whether a passenger is intoxicated or exhausted would not result in the fare being denied, but would rather lead to the passenger being paired with a more experienced driver, Uber says. Based on this, passengers could be matched with drivers "with relevant experience or training".