Imagine this: you and your friends relaxing on a couch, watching TV and enjoying a nice dinner. Now imagine doing this while being driven safely and quietly to a vacation destination. Cars that can drive themselves without the need for a driver, or autonomous cars, are getting closer and closer to becoming a reality, and the benefits seem to be revolutionary: reductions in travel time and traffic, cheaper transportation for the masses and more mobility for the impaired and elderly. The biggest possible benefit from these cars would be the improvement of safety: “It is estimated that [autonomous cars] could potentially address 94 percent of crashes involving unimpaired drivers" (Foxx qtd. in Chaitin). That is a very large reduction considering that motor vehicle accidents cause over 33,000 deaths per year in the U.S. alone (Centers for Disease Control).  This technology, previously only seen in futuristic movies, is now closer to reality than ever – the technology itself is currently being developed and even tested on real world streets. At the rate the technology is advancing, fully autonomous cars look as if they could hit the consumer market in just the next few years. However, with the amount of scientific research into autonomous cars, it is safe to assume that their availability to consumers is much farther away than recent advancements in the technology seem to suggest. There are many issues that may prolong the development of autonomous cars, ranging from the legal implications, such as liability in an accident and responsibility for software updates, to business interests from the companies developing the cars and insurance companies, and even the ethics within the software controlling the cars. 

Throughout my argument, when talking about fully-autonomous cars, the car must fall under SAE International’s level 5 for autonomous vehicles. SAE’s levels of autonomous vehicles range from 0 to 5; 0 being no automation at all with the driver in full control, and 5 being full automation with zero input needed from the driver (SAE). These levels have been adopted by the U.S. Department of Transportation as the official system for defining the level of automation a car has. Autonomous cars can “see” the outside environment through a vast array of cameras and sensors, such as video cameras, radar sensors, ultrasonic sensors, and LIDAR sensors. LIDAR (light detection and ranging) consists of a pulsing laser and a receiver, and can measure how far away something is by measuring the laser’s pulses. LIDAR can measure a smaller, more targeted area, compared to radar which has a broader field of view (What is LIDAR?). A central computer gathers data from the sensors and uses it to analyze road lines, signs, traffic lights, other cars, obstacles, and anything else that human drivers can pay attention to. The idea behind the safety of autonomous cars is that by having all cars controlled by complex driving algorithms, human error can be removed entirely, drastically reducing the overall number of accidents.

One of the biggest issues that will hold back autonomous cars is liability – more specifically, liability in an accident. Legal issues like this will take a fair bit of time for governments to develop a fair and effective system. Currently, with non-autonomous cars, when there is an accident, one person is usually deemed “at-fault” for the accident because of mistakes they have made, and they then take responsibility for repairs and damages. However, in the event of an accident involving fully-autonomous vehicles, there would not be anybody at the wheel to make a mistake; the cars are in full control. So, in this situation, who would take the blame? Who would be responsible for any damages and repairs? There are a few possible solutions that have been proposed, but they would all require major restructuring of the legal side of driving, all of which would delay autonomous vehicles from reaching consumers even more. One plausible idea is to put the responsibility in the hands of the manufacturers, “they are, after all ‘ultimately responsible for the ﬁnal product’: the vehicle including the system guiding it” (Marchant qtd. in Hevelke, 620). The biggest issue with this is that it would drastically hinder the development of the technology; why would manufacturers want to develop a product where they knowingly take full liability which can drastically hurt them? (620). Another possible solution is making the owner or operator liable, but require them to always be paying attention to the road and be ready to intervene if the car could not avoid an accident. While this does solve liability without too much restructuring of current laws, this solution has a major issue: it completely defeats the purpose of an autonomous car. An autonomous car is supposed to not only allow the occupants to relax and let the car itself without any input, but also have reactions much quicker than humans and not react in a panicked fashion. Forcing the driver to take over in such an event would rely on human reaction times rather than computer reaction times. Human reaction times are much slower than computers’, and this would cancel out the improved safety that autonomous cars promise (623-626). The issue here calls for a more in-depth solution, a process which will take years to research, write, revise and pass. The technology is already advancing faster than the law can keep up with, “Everything is racing ahead of a regulatory structure ill-equipped to usher in this change.” (Davies). Autonomous technology will be ready for consumers long before the government is ready for it.

The public’s attitude towards autonomous cars also will play a role in their implementation, because if they do not receive support or receive negative feedback, there is a chance they will not catch on. A recent survey from Deloitte, a U.S. based consultancy firm, says that 74% of Americans do not think that autonomous vehicles will be safe when they are released for the consumer market (Schiller). The distrust among the public has not been helped by recent fatal accidents involving Tesla’s Autopilot; a semi-autonomous system that can take over throttle, brakes and steering on highways and interstates. In one accident, a tractor trailer turned in front of a Tesla with Autopilot on. The system failed to detect the trailer, mistaking it for an overhead road sign, and so it did nothing to slow down or void the accident. The driver was killed as the car hit the side of the truck’s trailer, continued under it and off the side of the road. “The fact that the Autopilot system didn’t detect the trailer as an obstacle prompting emergency braking or steering is what is worrying a lot of people” says Fred Lambert from Electrik, a news publication dedicated to the transition to electric transportation (Lambert). 

Not only will the process of writing new legislature for autonomous cars be lengthy time-wise, the legislature itself could “could create more problems than they solve” (Smith qtd. in Davies). Autonomous technology will influence many different aspects of life, so implementing new laws could have negative effects in other aspects: “After all, this evolving technology permeates so many parts of society… [for example,] a law aimed at cutting congestion could tangle with tort law” (Davies). Davies details that certain laws, while having good intentions, may negatively hurt other aspects of law regarding the autonomous cars. Writing legislature that will not negatively affect these other aspects will be difficult to create, therefore adding more time to the already lengthy process of creating laws for autonomous cars. But even more than that, an issue for manufacturers is the possibility of individual states creating different regulations that all together makeup a “patchwork” of different regulations. This could make it very difficult for manufacturers to build cars and systems that comply with all the different states’ regulations. “That’s why the GMs and Googles of the world see federal intervention as a potential savior…[it] provides a broad, consistent framework for testing and deploying their robots” (Davies). 

One of the more interesting roadblocks for autonomous cars deals with the ethics of implementing autonomous cars. Patrick Lin from Wired demonstrates one of the many ethical issues faced by autonomous cars. For example, if autonomous cars 16,000 lives annually, “we really mean that they can save a net total of 16,000 lives a year” (Lin). The reduction in deaths is not just saving 16,000 people, for example it could be saving 20,000 people yet killing 4,000 new, different people than would have been killed before autonomous cars. “Some current non-victims — people who already exist — would become future victims, and this is clearly bad” (Lin). This “trading” of lives could prove to be a problem with the legislation being written. Another issue of ethics is the car’s decision of what to do if a crash is inevitable. If an autonomous car with 5 passengers is faced with a situation in which it crashes into an oncoming truck, killing all the passengers or swerving off the side of the road killing an elderly woman, which would it choose? (Nyholm et al., 4). Nyholm asks, “Should the car be programmed to always prioritize the safety of its passengers?” Or should the car be programmed so save the most lives, or maybe to save the innocent? (4). This ethical dilemma is a major hurdle the developers of autonomous cars (as well as government officials involved in writing the legislature for this technology) must clear before the public can enjoy be driven around town.

However, many believe that we could see cars without drivers on the street sooner than predicted, within the next 3 or 4 years even. The self-driving cars, much like most consumer-electronics, will be able to receive software updates (McMahon). This has the potential of getting the autonomous vehicles on the road sooner, as manufacturers could possible sell cars to customers while they continue to develop the technology. As the engineers and programmers fix bugs or add new safety features, they will can just push the updates out to consumers’ vehicles which would then automatically install, not requiring the owner take their car to a dealership to have it updated. Forbes Magazine uses an example of this from Adam Thierer, who is a researcher at George Mason University. Thierer explains of a scenario where a cyclist in front of a Google autonomous car was at a stop light, but instead of putting a foot down to balance, he could completely stand on the pedals, rocking back and forth to stay balanced without moving. He said, “the Google driverless cars were seeing people [doing this], and they didn’t know what to do, because it looked like he was going, he was stopping, he was going, he was stopping” (Thierer qtd. in McMahon). Thierer pointed out that Google engineers had to program the vehicles to recognize this and to know what to do, and then the updates were pushed to the vehicles (McMahon). This, however, does nothing to solve the large issues of ethical decisions, liabilities and insurance which are crucial to solve before the cars hit the market. That leads into another reason why some believe the cars may be ready sooner than expected: government officials in the UK have begun to create the legislature necessary for autonomous vehicles to be road-legal. The proposed bill includes many changes and requirements, including requiring more gas stations in the country to install electric charging stations, working on modifying insurance coverage, and even partially tackling the liability in an accident. For insurance, the bill says that there will need to be coverage for “when the driver is in manual control and when the car is driving itself” (Brian). This allows innocent victims involved in accidents with autonomous vehicles to get compensation for damages and injuries through the owners’ insurance, rather than having to wait to get compensation from the manufacturers. Most importantly in this bill is the liability covered. The bill states that if the owner “decided to modify the software on their vehicle or has failed to install important updates when their policy instructs them to, they will be made liable for any damages” (Brian). This will encourage users to update their autonomous vehicles with the latest bug fixes and safety improvements. 

Another thing that may significantly speed up the development of the technology is Department of Transportation’s (DOT) decision to begin building 10 autonomous vehicle test tracks, available to manufacturers needing to test their systems. The benefits of testing on closed tracks include the ability to create and repeat new situations, cover more miles and eliminate the risk of an accident with someone on public roads. But the biggest possible advantage of the tracks is what the DOT may require of manufacturers to them gain access to the tracks. The DOT may require users of the track to share their data with all the other users of the track. This could shorten the development time of autonomous cars by forcing manufacturers to collaborate and all build upon a collective set of data and research, therefore progressing at a much faster rate than if the individual companies were to all do the same research on their own. However, what causes this idea to fall through is the DOT’s requirement to share findings. Large manufacturers may not want to share their research – doing so would level playing field, no company would be able to have an advantage software-wise. This could potentially hurt the companies from a business standpoint, making them not want to use the test tracks at all – which would negate any increase of the rate at which the autonomous cars 

Fully autonomous vehicles are quite literally going to change the world; making it safer, greener, and more efficient – but as explained in this paper, not as soon as many are expecting. The technology itself is advancing at astonishing rates, with it being expected to be functional within next few years. Many believe that as soon as the technology is functional it will be released for consumer markets, but large barriers still block the path for autonomous cars, such as the legal problems of liability and the very complex issue of ethical decisions the cars will make. These issues, until fully addressed and solved, will prevent the sale of the self-driving cars, no matter how safe they prove to be. There have been different ideas and solutions proposed for most of the issues I have presented, but not only do they not cover all aspects of the individual issue, a single plan that completely solves every aspect of all the issues without interfering with other laws and without hurting the consumers, the manufacturers, or anyone else involved with the autonomous cars. The coming years will be crucial for manufacturers as well: with governments like the UK already proposing bills and laws to regulate the new technology, the manufacturers will have make sure they conform to the regulations set up, and are able to pass any safety tests and trials that may be required before they are approved to go on sale. 
