The application, "Predicting user state using machine learning", is how Uber detailed the promised technology.
The behavior of the person ordering the rise will be compared to their "normal" behavior - that is, it will be compared to long-term use patterns. The system relies on an algorithm to weigh a variety of factors, including typos, how precisely a user clicks on links and buttons, walking speed, and how long it takes to request a ride.
It is not known if the info will be used to deter drivers from picking up such passengers in order to sidestep potential trouble. For those who appear particularly intoxicated, drivers with special training or experience in dealing with rowdy passengers might be sent. As for riders, they may not be able to opt for shared rides if they're too inebriated.
The concern is that people deemed excessively drunk might be refused service from Uber, which could lead to them trying to drive home themselves or getting into other unsafe situations.
The patent application's authors are current or former members of Uber's Trust & Safety team, which works to make the company's products safer. If it does, you might end up having to call friends for a lift home at the end of a heavy night.
- 'Mr. Kim!' DPRK Leader Loudly Celebrated on Night Out in Singapore
- Salmonella outbreak linked to pre-cut melon — CDC
- Amazon expands Whole Foods discounts to North Carolina and 9 other states
- Luka Doncic staying in 2018 National Basketball Association draft
- Trump Is a ‘Pathetic Little Man-Child,' Says Former Trudeau Policy Adviser
- Ubisoft Shares New Screenshots For Tom Clancy's The Division 2
- North Korea media breaks tradition to report summit
- Starbucks outlets will close this afternoon
- Giant Martian Dust Storm Threatens Opportunity Rover
- Neymar Slams Austria's 'UFC' Tactics After Brazil Win