Uber’s popularity has been growing enormously since it began, as a safe way for people to get around. It is often used post-night out by people as a safe way of getting home and not drink-driving.

Their growing interest in drunk passengers has been broadcast this week by filing for a patent at the USPTO for a system that would use machine learning to determine the ‘state’ of a passenger.

Uber is really interested in detecting the difference between users of sound mind and users who are under the influence.

CNN first spotted the patent, which describes a method of measuring the user’s behavior on their phone against their usual behavior, using information like location, data input accuracy, data input speed, interface interaction behavior, the angle at which the user is holding their device or even the speed at which they’re walking.

The patent also describes a system that would notify drivers of the passenger’s ‘state’, theoretically letting them prepare for the adventure ahead. The patent says that riders in a particularly unusual state may be matched with drivers who have special training or expertise, or may not be provided service at all.

Of course, only a fraction of a company’s patents ever make it into the final product. Only time will tell if Uber’s idea to monitor the state of passengers will end up in the app.