Lentis/Self-Driving Cars

Introduction

Self-driving cars are increasingly in the media spotlight. Cars like the Tesla Model S and Google Koala make more and more headlines as more consumers purchase them. Their abilities to operate without a driver offer many perks. However, many have raised concerns over how the cars affect the people who own them and people who don't.

Automation

A key distinction to make with regards to self-driving cars is the difference between self-driving and driver-less. Driver-less cars are actually a subset of self-driving cars while the self-driving cars have varying levels of autonomy. There are 5 different levels of autonomy as classified by SAE international, six if you count no autonomy. At level one, the car is controlling either the lateral or longitudinal movement. An example of this is any car with cruise control. With cruise control, the car is controlling the speed while a driver is steering the car left and right. At level two, the car can take control over all aspects of specific scenarios. For this level, parking assist is a prime example. At level three, the driver no longer has to monitor the system at all times. However, he must be ready to take over at a moment’s notice. At level four, the car is fully autonomous for most purposes. A driver only needs to intervene in extreme cases such as severe storms. A car that reaches level five autonomy attains the pinnacle of self-driving. All a driver is required to do is to enter a destination and the car does the rest.

Autonomy Level	Characteristics
0	No autonomy
1	System controls either longitudinal or lateral movement
2	System fully controls car in specific situations. Driver must monitor system at all times.
3	System fully controls car in specific situations. Driver can divert attention from driving.
4	System controls car in nearly all situations. Driver intervention only needed in extreme cases like severe weather.
5	System can cope with all situations. No driver needed

Early Self-Driving Cars

Firebird II

Experiments with self-driving cars began in the early 1900's. These early designs focused primarily on automating small aspects of the driving experience. The American Wonder was one of the earliest forays into self-driving cars. While not technically autonomous, the car shows that the field began early in the 1920s. The designed fitted a radio receiver to a car and remotely controlled it in a following car that had an operator with a radio transceiver. Another example is the Firebird II, manufactured by General Motors. This car was equipped with an experimental form of driving assist that allowed the car to stay within its lane. The system relied upon an electrical wired that would be embedded underground and send information to the car.

Google Koala

Finally, the Google car is another, more modern example. Developed under GoogleX, the Google driverless car began as a refitted Toyota Prius in 2009. The system was tested primarily on secluded roads with little traffic. After 300,000 miles of testing, a Lexus SUV was refitted for driving on freeways and city streets. Testing begin with the SUVs in 2012. In 2014, Google produced the Koala. An important feature of the Koala is that it lacks a steering wheel and is fully autonomous. Furthermore, the Koala was the first car where there was no technician present in each car to intervene if something went wrong.

Potential Benefits

Self-driving cars offer many potential advantages. The first few are related and stem from having a computer controlled car as opposed to a human operator. A computer is able to react much more quickly and should lead to improvements such as fewer collisions, less traffic congestion, and higher road speeds. The lessened traffic congestion will be a result of computers requiring a smaller safe following distance. This will allow more cars on the road with greater efficiency. Along the same lines, a computer will be better able to perform at a higher speed and still react appropriately compared to a human driver. Other advantages may be a more effective workforce. This will stem from commuters no longer having to focus on driving. They can focus their attention on fully waking up or preparing themselves for their work. A final advantage that will be mentioned is the possibility of car sharing. For this, a fully automated Uber service works as a fine example.

Challenges

Ethical Dilemma

The manifest function of self-driving cars is to eliminate human error, the main source of car accidents. Ethical dilemmas arise in situations where the car needs to make a calculated decision on what to do during an accident. These decisions will be shaped by the influence of self-driving car manufacturers, policy makers, and the drivers and passengers themselves.

Moral Machine

Moral Machine, a website created at the MIT Media Lab, aims to gather human perspectives on moral decisions made by self-driving cars. It presents scenarios where a self-driving car that has a sudden brake failure has two options: to continue ahead and drive through a pedestrian crossing, resulting in the injury or death of the pedestrians, or to swerve into a concrete barrier, resulting in the injury or death of its passengers. In these scenarios, victims vary in age, occupation, species, gender, and fitness. Data collected from the Moral Machine gives insight into the trade-offs that people will make in regards to self-driving cars. The Moral Machine also shows people how difficult it is to make these trade-offs, and helps them understand that this is the task that regulators are faced with, and the types of regulations that may be made in the future.

Social Dilemma

Self-driving cars pose not only an ethical dilemma, but also a social dilemma of public safety, and getting society to agree on what trade-offs to make and how to enforce them through regulation. The social dilemma of the tragedy of the commons is exemplified in this scenario: if all self-driving car owners want to protect themselves and their passengers at all costs in case of an accident, they are collectively diminishing public safety. However, in most cases it is not up to the owners to directly decide what their cars do during accidents. Regulation and law are reflective of the values of society, and consumers, car manufacturers, and regulators all play a role in solving this social dilemma.

Regulation

Social dilemmas are often solved by regulation, where a group decides for society as a whole what is best for them given the constraints, and how to monitor and enforce their decisions. In September 2016, the U.S. Department of Transportation released their Federal Automated Vehicles Policy Overview, which includes a 15-point Safety Assessment for self-driving car manufacturers to consider. One issue in the Safety Assessment list is "Ethical Considerations: How Vehicles are programmed to address conflict dilemmas on the road".

Social Implications

With new technology comes new issues that need to be addressed and examined. Self-driving cars can pose a threat to humans. There is a reoccurring social pattern with the integration of new revolutionary technology; as technology increases the convenience of individuals’ lives, they start to lose touch with simple abilities that they were capable of prior to their dependence on the new technology. Some examples of these technologies are global position systems and calculators. Before the GPS, individuals navigated the road using maps and reading signs. Now many individuals struggle with using those skill-sets and depend heavily on their GPSs to navigate roads. Before calculators, individuals were able to do mental math easily and quickly. Now even for simple quick calculations, many have to depend on their calculators. These technologies may be eroding individual’s senses and skills. There is a point where machine capabilities start to control us rather than empower us.

The need to practice driving skills will not be needed with the adoption of self driving cars. However, driving is the modern form of practicing human survival. When driving, individuals have to be spatially aware in order to get to their destination safely without harming others on the road. Driving requires individuals to make life-death decisions, forcing individuals to practice their situational judgement calls and improve their reaction times. Self-driving cars may make people dumber, less alert, and less involved with transportation. People’s sharpest senses will be dulled, making them more complacent and mentally thick since humans are submissive to the easy way out. Although self-driving cars increase productivity and help ease everyday life activities, it could increase human's laziness instead since these new technologies do not allow individual's to maximize their brain's potential.

Environmental Implications

Self-driving cars have both positive and negative impacts on the carbon intensity of transportation. The interactions of this technology with the environment manifest at vehicle level, transportation system level, urban system level, and society. Focusing on the vehicle-level improvements is likely to yield excessively optimistic estimates of environmental benefits.

Unemployment

There are numerous other important unintended consequences of self-driving cars that require examination. Unemployment will increase significantly since driving-related jobs will decrease. Taxi drivers, truck drivers, and delivery drivers will be forced out of their work. These jobs’ pay will also decrease since their high pay is due to the hazards of their job which are now reduced due to self-driving cars. Although the new technology will develop new jobs, the skill sets required for those positions are not easy to learn and require resources that many disadvantaged individuals don’t have access to. These higher-level jobs are also more easily off-shored while transportation jobs are not. Occupations in the transit services, car shops, and automobile insurance industries will also suffer.

Security

Self-driving cars pose numerous security dangers. Since these cars are dependent on software, security holes will always exist that can be exploited by hackers. Even if a way to exploit the software doesn't exist, hackers can always invent a way to exploit them. These systems also increase vulnerabilities to terrorism and natural disasters. Self-driving cars will most likely be controlled by the same satellite system; this centralization makes it more likely for things like natural disasters to put these networks offline especially when they are needed most. There will also be serious privacy concerns since the software will always know where the car is going or where it has been. Preserving the privacy of the user will be a challenge.

Conclusion

The development of self-driving cars is rapidly picking up in pace as many companies continue to compete. Looking at the cars manufactured by Tesla and Google, they are getting ever closer to becoming fully autonomous and no longer requiring a human operator. While this autonomy offers many potential benefits to consumers and society as a whole, there are many moral and social challenges that still need to be overcome. It is important that all of these issues be addressed in the coming years before self-driving cars become mainstream.