UBER and Self Driving Cars

Featured article by Jeremy Sutter, Independent Technology Author

In science fiction stories about future society, one of the mainstay tropes is the self driving car. Such books, movies and TV shows let us see corporate movers and shakers making deals, slackers playing video games and harried office workers putting the finishing touches on a project – all while the vehicle gets them safely and efficiently to where they are headed.

With the exponentially improving computer technology of today’s world, one of the horizons many companies are headed toward is the realization of this self driving vehicle dream.

How Close Are We to Having Self Driving Cars?

Strangely enough, both a long way off and closer than most people imagine.

We are a long way off because of the difference between how computers deal with the world and how a human driver deals with the world. First, computers cannot make associations that they have not been specifically programmed to make.

For example, a human driver is cruising through a residential neighborhood and sees a ball roll into the street. Knowing that there is a good chance a child will be chasing that ball, the human slows down and focuses on the side of the road where the ball originated to be able to react as quickly as possible. The computer would need to be programmed to recognize the ball, not just an object rolling into the street, but a ball. From a golf ball to a basketball and everything in between, the computer would need to be programmed to recognize each object and call the routine that tells it how to respond. This would need to be done for every imaginable object a child could throw, kick or roll into the street. While a human instinctively makes the decision that something is a toy, the computer would need to be told that a toy truck, a hula hoop and a soccer ball all require the same response.

In addition to programming specific responses to a vast array of different obstacles and situations, these things would also need to be programmed for any condition the self driving car might encounter. How difficult this is can be inferred from the US armed forces’ experiments with creating computerized scanners to automatically detect camouflaged vehicles. After programming the system, they tested it and it instantly spotted a vehicle that they had camouflaged nearby. Testing it again, it didn’t. Over many tests, it would sometimes spot the vehicle and sometimes it would not. The programmers could find nothing wrong. After a lot of time and careful observation of the tests, they eventually discovered that the times it was failing coincided with a cloud overhead putting the camouflaged vehicle in shadow. This difficulty was with a computer designed to do one very specific thing. Imagine how much more difficult it will be to correct unexpected computer reaction to every different condition a vehicle might encounter on a road in all seasons and at all times of the day.

In addition to the technological difficulty, there is the legal difficulty. Taking advantage of nebulous legal language in the Texas highway laws, Google recently performed a live test of self driving cars in Austin, bringing up the question of whether a computerized navigation system can be as safe as a human employing defensive driving. El Paso TX legislator Joe Pickett predicts that every mishap with self driving cars will be followed by a nationwide barrage of legislation related to that mishap.

In spite of these difficulties, one of the major players in the race to a true self driving car is UBER, a company that is actually using automated vehicles (with a human driver present to monitor) to taxi customers from place to place.

Why is UBER Making This Push?

UBER is losing money. Being forced to split fares with drivers is a huge cost for the business. They envision rolling out a fleet of automated taxis that bring in money with little downtime and fares that go entirely to the company.

How Could This Backfire?

There are two ways:

First, UBER would be making itself open to many huge lawsuits when, inevitably, their self driving vehicles are involved in accidents. This added liability is a frightening concept for a company that is already having a rough time.

Second, imagine that UBER manages to successfully create safe self driving vehicles that satisfy lawmakers enough to roll out a fleet of them. How far behind will the competition be? The company that finds success with a technological innovation is often not the one that developed the technology, but one that sits on the sidelines and watches the process and learns from any mistakes the trailblazer makes.

While it would undoubtedly be good for society at large to have access to safe self driving vehicles, it could be bad for UBER.