Are Autonomous Car Makers In A Hurry To Reach A Driverless Future?

An Uber self-driven car hit and killed a woman crossing a street in Arizona on Monday, March 19th. It was a Volvo XC90, a Swedish car owned by China’s Geely. While the car was in autonomous mode, there was a backup driver at the wheel. The car was reportedly traveling at 65 kmph and the accident happened at 10 pm. This is the first known pedestrian fatality involving a self-driving car.

Volvo quickly responded that the software being used in the vehicle was not developed by it. And Uber promptly removed all its self-driving cars from the roads. Quite predictably, this incident has revived the discussion on the how safe this technology is? Will we ever see cars with drivers getting replaced by driverless cars? How is the regulation going to cope up? Who is to be blamed if an accident happens? What about the ethical angle – when an algorithm-driven AI system makes a mistake?

It is true that no one asked car makers or tech companies to make cars driverless. And accidents like these bring up the question whether the tech giants and the techno-optimists want to force upon us their multi-billion-dollar experiments.

As this Washington Post article suggests, vehicle fatalities are measured in terms of vehicle miles traveled. In 2016, there were 1.18 fatalities for every 100 million miles. Waymo, the industry leader recently reported logging its 4 millionth mile of road travel, much of it in conditions which can be termed as favorable. Uber just reached 2 million miles.

But this is unfair comparison. First up, let us acknowledge that this technology is still at a nascent stage. Automobile and tech companies are still in the testing stage. Practically, no self-driving car will be on the road before 2020. And those which will come, will certainly not start on the crowded streets of Mumbai. The roads and the driving conditions will be chosen carefully and monitored closely.

But still, this incident should force us to pause and take a hard look at the premise.

Take the backup driver, for instance. A backup driver behind the wheels of an autonomous car is not much of an assistance when most needed. Humans suffer from attention deficiency, poor response time, poor decisions, tiredness, state of drunkenness, and worst, they can’t stay away from their mobile phones. The technology behind a driverless car does not suffer from any of these. So, in theory this system will be far superior and safer. But, if we enumerate all the scenarios which must be taught to the AI system, we will quickly realize it’s a tall ask and perhaps a never-ending task. Hence, to get the desired outcome, it will take time and millions of miles of training in all kinds of scenarios. And so, despite the hype created around driverless cars, the make-believe world of Hollywood movies and the larger-than-life image of the tech giants behind the projects, the promised future is not so near.

What about the regulation and policy making? Will this case set any precedent in drafting a sound policy? Who should we hold responsible in this case? Was it the AI system or the backup driver sitting behind the wheels; or is it Uber who should be sued? Did the backup driver see the pedestrian, and how long before he did anything? When did the sensors detect the pedestrian crossing the road or did they malfunction? Did the AI system have enough time to respond? Was the distance enough for the system to have activated the emergency brakes and avoid the accident? Did the lighting on the street matter? Was the pedestrian at fault – how much?

The sheer volume of data which needs to be studied to come to an informed conclusion can be mind-boggling. How do you modernize outdated regulations in the face of fast-paced changes in technology and innovations? Is it now time that there is a central framework to drive this automation going forward – or will it thwart the innovators from innovating? Regulators across the world are still grappling with these questions. The real premise on which this is progressing is: ‘we don’t know the technology and the law will evolve’. In that case, how does society and the legal system deal with such accidents? Much is at stake.

Then there is the ethical question. How is death from an autonomous car accident different from one caused due to negligence of a human driver. The ethical problem is also often presented in the form of an idealized, unsolvable decision-making problem, called the trolley problem. Although, there are no ethically right answers to the problem, it poses an important challenge for an AI system to come up with an answer. A computer algorithm will solve it by running some rules; but will the outcome achieve a level of acceptable social norm? Which social norm do we take as a reference, since it would vary from society to society. In Uber’s case, who feels ethically responsible for the accident?

Perhaps autonomous cars can work when we accept the idea of that there will be no perfectly autonomous cars. It may need someone, maybe a government or a celebrity techno-evangelist to come out and say this and prevent a race to the bottom to put these cars on the roads faster than necessary. Maybe, the innovation should happen on the terms of the regulator rather than on the car maker’s terms.

What do you think?

----------

Enjoyed this article? Share it with your network.

----------

In case you missed

Amazon Wants To Open Checking Accounts. What Does It Mean For The Banks?

How is Artificial Intelligence Changing Our World? (part 2)

Getting Intelligent About AI (part 1)

I invite you to browse all my LinkedIn articles archived here.

----------

About the author

Anindya Karmakar has led multiple initiatives at the cutting edge of digital connectivity, IoT, robotics, AI, analytics, paperless branch and remote advisory. He is passionate about the digital revolution which is underway. He simplifies and de-clutters digital jargons and concepts and presents them in layman's language.

----------

The views and opinions expressed or implied herein are my own and do not reflect those of my employer, who shall not be liable for any action that may result as a consequence of my views and opinions. The pictures used have been taken from the open internet and I don't claim any credit for them.