Driverless Car

The future of driverless cars

2018.04

The future of driverless cars: Are autonomous vehicles capable of becoming drivers with morals?
Driverless Car
리스트 보기
인사이트
토론질문
영상/기사
국문보기
영문보기

I. The push for the mass adoption of driverless cars

The push for the mass adoption of driverless cars is happening rapidly and on a global scale

  • It is not uncommon for pedestrians of major U.S. cities to see tests of driverless cars as big companies have been conducting them on public roads.
  • The testing of “driverless delivery vehicles” that encompass testing for self-driving taxis and trucks are also actively underway.

Global technology firms and car manufacturers are moving aggressively to make investments and each carve out their share of the autonomous vehicle market, using their own core capabilities.

  • Google: Google began a self-driving car project with labs at Carnegie Mellon University and Stanford University in 2009. The project’s vision was to ‘fundamentally innovate driving as to achieve increased safety on the road, safer multitasking for riders, and a reduction of carbon dioxide emissions.’ Google’s self-driving technology has a track record of 3 million kilometers of testing on the road and is safe enough that a portion of Google’s employees use driverless cars to and fro from work (they drive as usual from home to the highway, but activate autonomous driving on the Silicon Valley highway). Just as Google replaced Samsung and Nokia as the leader of the smartphone market, it aims to [1]wrest control of the automobile market from Mercedes Benz. It plans to do so using massive amounts of transportation data, world-class engineering talent, and various applications including Google Maps that it already owns.
  • Uber: Uber is the key contender to Google in a battle for the best autonomous vehicle technology. Uber established its Pittsburgh-based autonomous vehicle lab in 2015 through the mass recruitment of key research talent from Carnegie Mellon University. Its founding vision is “to provide taxi services that are less expensive than bus rides through driverless cars.” Uber proceeded to acquire the self-driving truck startup Ottomoto in 2016 for a reported KRW 750 billion ($680 million) along with its key talent. Ottomoto’s key talent included not only, ex-Google, Tesla, and Apple employees who are experts in the autonomous vehicle technology space, but also Anthony Levandowski, the man who started Google’s self-driving car project. Uber’s application of Ottomoto’s technology has sparked a blockbuster legal battle with Google. Uber, which tests its driverless taxis on the streets of Pittsburgh, is the fastest growing player in the ‘driverless delivery vehicles’ space.
  • Tesla: Tesla, known for its successful commercialization of the world’s most luxurious and technologically advanced electric cars, is also a key contender. It is developing autonomous vehicle technology with the goal of manufacturing ‘technologically flawless driverless electronic vehicles, opening the door to a new era of commercialized electronic and driverless cars before 2020’. Tesla’s autonomous vehicle technology is advanced to a point where all owners of its Model S have access to partially automated driving (e.g. Autonomous Emergency Breaking technology that prevents collisions). However, Tesla’s focus is not on developing technology for driverless delivery vehicles, but on developing technology for fully commercialized cars that maximize safety on the road and allows for full multitasking for its drivers.
  • Carmakers like BMW, Mercedes Benz: Carmakers like BMW and Mercedes Benz are late to the race, but quickly rushing to develop their own autonomous vehicle technologies. The push comes after realizing that tech firms developing driverless car technology will have immense effects on the automobile market in the long term. BMW is aggressively investing in the technology, announcing that it plans to “release a driverless car by 2020 as part of a car-sharing services.” The American multinational carmaker GM is less focused on the commercialization of driverless cars and more on developing original technology that it can license. GM stands out in that it has racked up more autonomous vehicle technology patents than Google or Toyota.

Policymakers and governments all around the world are introducing both regulations and supportive measures for driverless cars. The U.S. has introduced laws in certain states that fully support development and testing, leading the way for the mass adoption of driverless cars.

  • Since 2011, the U.S. has provided legal backing for autonomous vehicle testing for companies in certain states.
    • Egged on by powerful lobbying from Google, the United States government (for the first time ever) authorized the operation autonomous vehicles in Nevada in June of 2011. The granting of legal operation of autonomous vehicles in Florida and California followed suit.
    • In 2016, the National Highway Traffic Safety Administration (NHTSA) ruled that artificially intelligent software of autonomous vehicles can be interpreted as “drivers,” a move that accelerated the mass adoption of driverless cars.
  • The U.S. Department of Transportation (DOT) announced that it will invest $4 billion over 10 years to support regulatory reform and R&D of autonomous vehicles
  • The NHTSA has categorized driverless cars into four distinct levels, so that it may enact policies for driverless cars that correspond to the level of autonomous driving.
    • Level 1: Partial automation
    • Level 2: Limited self-driving (“hands-on”)
    • Level 3: “hands off”
    • Level 4: “eyes off”

In conclusion, the self-driving car market is currently centered around the United States, with companies that own advanced autonomous vehicle technologies like Google and Uber leading the way.

  • Google and Uber, with their wealth of data and top engineering talent, are taking over the autonomous vehicle market with growth based on cutting-edge autonomous vehicle technologies.
  • The U.S. government is relaxing regulations on autonomous vehicle operations for Google and Uber, thereby supporting them on this quest.


II. The expected benefits and drawbacks of driverless cars

The biggest reason governments and large companies around the world are rushing to develop driverless cars is that their mass adoption will not only bring about market opportunities, but also a range of societal benefits.

  • Market opportunities: According to a forecast by Fortune Magazine, most of the vehicles on the road will be replaced by driverless cars by 2050. The market size for driverless cars is estimated to reach $7 trillion by 2050, with the sale of driverless cars alone accounting for $4 trillion. The remaining $3 trillion is estimated to come from the economic activity behind delivery and logistics.
  • Increased safety: When driverless cars become mainstream, vehicle crashes due to human error such as DUI and falling asleep at the wheel will be eliminated, as the autonomous car’s immediate response time and wide field of vision will reduce accidents (Note: 95% of all vehicle crashes are due to human error). Casualties and deaths from vehicle crashes will sharply decline, bringing with it a huge reduction of societal costs.
  • Reduced traffic on the road: According to many studies, the average number of passengers (aside from the driver) for all vehicles on the road is 2. Even 80% of all taxi rides consist of rides with 2 passengers or less, which means that by the time driverless cars become mainstream, most vehicles will be replaced by 2-Seater cars. When 2-Seater cars become the norm, traffic problems are expected to go down by 40% and lead to positive [2]ripple effects. Reduced air pollution will lead to heightened quality of life for people everywhere.
  • Reduced transportation costs: Mass adoption of driverless taxis will lead to taxi ride fares falling below that of bus rides. The cost reduction is expected to come from inside-taxi advertising and lowered salaries for taxi drivers. At the same time, mainstream adoption of driverless trucks will drive down the cost of logistics by 30%.

⇒ The mainstream uses of driverless cars are expected to lead to a sharp decrease in pollution, traffic, and transportation costs. It will also increase safety on the road and the quality of life for people as a whole.



However, there are just as many risks and concerns to address:

  • The possibility of large-scale car accidents: Bugs and viruses in the AV technology can lead to critical system errors, which may cause large-scale car accidents
  • The legal question of blame: Accidents and crashes of driverless cars introduce the complicated legal question of who is at fault.
  • Engineering morality: in cases where driverless cars need to make a moral decision on the road, the issue of how to engineer such decision-making capabilities arises. For instance, in the rare case that a car has to crash into either an old lady or a child, the artificially intelligent machine will be tasked with making that moral decision.
  • Ethical dilemma: When driving as usual, the burden falls on the human driver of the vehicle and he or she takes full responsibility for the decision that he makes. In the case of artificially intelligent vehicles, these hard decisions must be engineered through algorithms and lines of code, which introduces the further question of who decides the moral standards that will govern these machines.

There are also signs that errors and selfishness of key decision-makers in the path to mainstream adoption of driverless cars will cause big problems.

  • In the race to become the leader of the driverless car market, governments and companies may ignore safety requirements when it comes to testing autonomous vehicles on the road. Bypassing cautionary measures could lead to casualties and deaths.
  • The many systems and policies that will be erected in the world with driverless cars may be engineered to benefit a particular group(s).

Driverless car accidents have already happened and all around the world, and they are raising the question about the safety of autonomous vehicles.

  • Tesla’s driverless car accident: In May of 2016, a driver was killed while driving a Tesla Model S on autopilot mode (which is driver assistance mode, not full autonomous driving) when the car collided with a semitrailer truck. The deadly crash happened because the Model S did not detect the white trailer truck that cut in front of the car, which would have used its breaks to stop. Instead, the car kept its course and crashed into the truck, which was making a left turn on the highway intersection. The police statement said that the Model S’s vehicle detection system failed and could not distinguish between the bright sky and the white colored truck. However, the statement concluded that the deceased driver was at fault for not having his hands on the wheel in spite of the many alerts that the car sounded. In March of 2018, a driver who was driving a Model X on autopilot mode was also killed in a fatal crash. The autopilot malfunctioned and crashed directly into a barrier, caught fire and got hit by two other cars. The police are still looking into the cause of the crash and whether the driver’s negligence or the autopilot technology is at fault.
  • Google’s driverless car accident: In Mountain View on September of 2016, a truck ran through a red light and collided with one of Google’s self-driving cars. There were no casualties, but the side of the car was completely crushed. Google issued a statement saying, “in accidents involving Google’s self-driving cars, it has almost always been the other driver who was at fault” and added, “our task at hand is developing an autonomous vehicle technology that can react appropriately to other unpredictable drivers on the road.”
  • Uber’s driverless car accident: On March 18th, 2018, Uber’s self-driving Volvo struck and killed a woman crossing the road at night, not at a crosswalk. The car was traveling at 60 kilometers per hour with an Uber employee at the wheel, but the car did not detect the person crossing the road in the dark. The authorities issued a statement saying “it would have been difficult for even a human driver under the same circumstances to avoid an accident.” However, after video footage of the accident was released many are blaming the vehicle’s detection technology for the fatal accident. Although the victim was not at a crosswalk, many say that the technology, using its camera, sensors, and radars, still should have stopped the vehicle. They hold that there probably was a bug in the technology and Uber is to blame, since its self-driving car was not capable of responding to unpredictable situations such as this one.


III. How to mitigate the risks and drawbacks of driverless cars

The push to mitigate the drawbacks of driverless cars are happening on technological, legal, and ethical levels.

  • Technological efforts: Companies are intensively testing their autonomous vehicle technologies (through an application of AI, machine learning, GPS, and state of the art sensors) in order to develop “fully autonomous vehicles that require no human intervention with an ability to function, even under highly stressful conditions.”
  • Legal efforts: Governments from around the world are scrutinizing the autonomous vehicle technologies of relevant companies and creating policies tailored to each situation. Certain states in the U.S. have set up autonomous vehicle technology-friendly policies and regulations for firms in the driverless car industry (and those looking to get investment to enter the industry). As a result, active autonomous vehicle testing in these states has led to the rapid advancement of the AV technology.
  • Ethical efforts: Engineers and Stanford University and MIT are using AI, machine learning methods and rigorous autonomous vehicle testing to develop self-driving cars that can make value-based ethical decisions (as humans do) when faced with moral dilemmas. For instance, research labs amass large amounts of data on real ethical decisions humans make when faced with moral dilemmas and use that data to train machines. (e.g. labs ask a sample of the population ‘which person would you crash into if you had no other choice? Why?’ and train machines through datafication of these answers). Many have raised concerns that this method may make driverless cars to make largely utilitarian decisions.

In spite of a [3]multifaceted approach to mitigate problems, concerns about the problems that come with the mass adoption of driverless cars still loom large.

  • What should be the moral standard by which autonomous vehicles make decisions? Is it even right to have artificially intelligent cars, that cannot take responsibility for their actions, make value-based decisions?
  • What should we do about all of the taxi drivers and truck drivers who will lose their jobs with the mass adoption of driverless cars?
  • Is there such a thing as a 100% full-proof technology? In a world that is already heavily (and increasingly) dependent on machines, a single technical error can have huge consequences. Do we have the human intelligence and morals to prevent and control for disasters?

Once driverless cars become mainstream, it is clear that the number of vehicle crashes and traffic problems will decline. However, it is also clear that getting there will take much [4]trial-and-error and inevitably bring forth conflict, not to mention negative externalities.

What are your thoughts on the future of driverless cars?

또는
이메일로 회원가입
이름 *
이메일 *
비밀번호 *
비밀번호 재입력 *
추가 정보 입력(선택)
제휴회사
추천인