press enter to search

Patent suggests AI learning could be used to identify drunk Uber passengers

Anthony Wallace

Agence France-Presse

| Tue, June 12, 2018 | 04:01 pm
Patent suggests AI learning could be used to identify drunk Uber passengers

CNN spotted a patent filed on June 7 with the United States Patent & Trademark Office, titled ‘Predicting user state using machine learning' in order to alter the trip parameters.' According to CNN, the patent application's authors are current or former members of Uber's Trust and Safety team, however the company has not commented on the patent. (Shutterstock/File)

Uber may implement a system that determines if users are behaving like their usual selves.

Uber may want to give their drivers more complete information as to how intoxicated their passengers are, if they are sound of mind and allow drivers to compare behavior from one ride to another, according to a patent filed recently.

CNN spotted a patent filed on June 7 with the United States Patent & Trademark Office, titled ‘Predicting user state using machine learning' in order to alter the trip parameters.' According to CNN, the patent application's authors are current or former members of Uber's Trust and Safety team, however the company has not commented on the patent.

The content of the filing suggests that the popular transport app that connects drivers with passengers wants to implement a system that would try to determine a user's state upon requesting an Uber and reference it against prior behavior or characteristics. Although it isn't exactly clear how the platform would do this, it does suggest an assessment of how quickly and accurately the user can input data, how fast they're walking and the way they interact with the interface.

Read also: Lesson from Go-Jek: Answer to Uber, Grab

Where and what time the request is made could also help paint a picture, for example, a user making typos and walking slowly through a bar district at 4am has probably had a few drinks.

“Safety incidents and personal conflict incidents, can occasionally occur when users and/or providers behave uncharacteristically,” says the patent.

So in an attempt to avoid negative transport experiences for both user and driver, Uber could use this information to determine whether a more experienced or trained driver should be assigned to the user. Furthermore, it could prevent the user from joining a trip with other riders, for example with Uber Pool.

One example the patent provides suggests that a user may be “uncharacteristically tired when requesting a trip” and therefore could have difficulty locating the Uber, so it would be “desirable to minimize impact of such safety incidents in travel coordination systems.”

Comments