A patent has been filed by ride-hailing service Uber for technology that would use machine learning to determine whether or not a passenger is inebriated.

Designed to detect passengers behaving “abnormally,” the patent doesn’t go into detail about an exact definition of what abnormal is, yet there are suggestions that the technology would be used to spot signs of drunkenness and fatigue.

That, coupled with details such as how the person using the app is walking, the speed of their walk, how they are using their app, their typing speed, accuracy and phone angle, in addition to where and when they order a ride, can all be fed into the algorithm.

Once identified as an inebriated passenger, certain steps might be taken by Uber, including pairing the passenger with a specific driver or giving the driver a head’s up about the condition of their passenger.

"It would be cool if drivers got extra money for picking up drunk passengers. It's not a big deal, but it definitely gets old after about the 25th time," wrote Harry Campbell, author of The RideShare Guy blog.

The purpose of the algorithm is to make rides safer for the both the passenger and the driver, as well as heading off potential personal conflicts between driver and passenger when the passenger is behaving abnormally.

This patent comes on the heels of reports that for over the last four years, 103 Uber drivers have been accused of sexual assault — with many of the alleged occurrences happening when the passenger had too much to drink.

Likewise, the technology might help protect drivers from intoxicated passengers who become abusive.