Recent advances in technology have allowed for a much faster rate of development for driverless cars. These leaps in progress have led to a shifting perception. Regardless of the growing push for technology, there are still a large number of ethical issues that autonomous cars currently face. If these are not solved or at the very least considered, there could be major ramifications for society, industry and the government in the near future.
Autonomous vehicles should not be widely introduced into society
Taxi, vehicle-rental and transit companies are openly planning to replace human drivers once it’s feasible. A Columbia University study suggested that with a fleet of 9,000 AVs, Uber could upend every taxicab in New York City. AVs do come at a cost: fewer transport-dependent jobs. It seems a recapitulating trend of technology advances broadcast the enthusiasts buying Google cars but is inured to the workers shed in hundreds of thousands. Nearly 4 million vehicle-based jobs; for schools, cities, or long-distance travel, are at risk by the next decade for the transport industry. Is this a trade-off actually worth making? Are the societal benefits holistically greater with automation of transportation labour concurrently eliminating jobs? Or are the well-paid jobs in transportation worth retaining as an essential pathway to the middle class?
Considering the current 40 million conventional vehicles road users, there exists the difficulty of road sharing between autonomous cars and conventional cars. AVs will have to contend with humans on the road with both parties having a different sense of judgement. AVs are designed to use machine learning from different driving scenarios but traffic mixing with humans requires both natural intuition and the ability to improvise in distinct scenarios. A study by UMTRI revealed that with traffic mixing with humans, driverless cars have a “higher crash rate per million miles travelled than conventional vehicles” although the accidents were not the fault of the driverless cars. Is it right to risk injury and life by placing both types of cars on the road?
Although, the emergence of driverless cars could mitigate 90% of traffic accidents caused by human error, they may pose a higher risk of disastrous accidents caused by technology failure or third parties (hackers) with the intention of causing road chaos. Advanced testing on technology can increase its reliability to a high degree but technology can be fallible and unpredictable. Google’s autonomous vehicles have already experienced 272 technology failures over 424000 miles during test-driving due to strange sensors readings, communication failure or problems in monitoring systems.
If the Trolley problem is considered in the context of driverless cars, some serious questions must be answered. If the car is involved in an unavoidable accident, should the car protect the driver, or sacrifice him/her? This action can have major ramifications, and furthermore asks if programs and companies should have the power to make these decisions. It seems rather strange to place our lives in the hands of a computer, as our ability to judge a situation is an invaluable skill that we possess, and utilise fairly well. Would consumers and the public be comfortable using a driverless car, knowing full well that they may be ‘sacrificed’ to preserve the life of another?
Autonomous vehicles should be more widely introduced into society
Driverless cars will definitely not exhibit the same driving style as regular cars, however there exists an opportunity to create new laws, rules and infrastructure surrounding driverless cars. Therefore, governments can create moral and ethical laws from first principle regarding driverless cars. Hopefully these laws will be rational and align with the general public’s moral compass. For example it would be silly and dangerous to other motorists to mandate that driverless cars must follow the law to the letter. Driverless car lanes or mandated behaviour in certain situations are rational outcomes.
Similarly to a new driver on the roads, most judgements and actions taken will be wrong. However the algorithm will optimise its internal machine learning parameters, just as a new driver would learn. Google for example has over 2 million kilometres of test data for thousands of driving scenarios from which algorithms can learn. This allows driverless cars to correctly identify a wide variety of scenarios, and reduces the likelihood of ‘un-human’ driving characteristics. For autonomous cars to be practical they only need to be better than humans, and in a way they already are. Driverless cars don’t get distracted, fatigued or overconfident. People can also develop poor driving skills, often as a consequence of bad training or habits. This can be just as detrimental as bad algorithms.
The public perception that autonomous vehicles are always required to make the right ethical judgement is flawed. Countering the Trolley problem, by definition there is no ethically correct answer. Humans would not always be able to make the right ethical choice either, so why should robots be held to a different standard? The probability of such an oddly specific dilemma occurring is extremely low, so practically speaking it is not worth investigating. Furthermore, 75% of respondents agreed with the utilitarian view of preserving life, showing that the general public’s opinion aligns with the logical decision argument. Self-driving cars do not need to make the correct decision; they just need to not make the wrong one.
Technology causes changes in society, and the introduction of the driverless car is no difference. As history has shown time and time again technology will replace the mundane and repetitive tasks that humans have been doing, it is an inevitable fact. Take farming as an example. In the U.S, 90% of the population used to be involved in farming, now only 2.6% of the American population are involved in farming. The invention of the tractor and increases in productivity has allowed an increase in quality of life. The same thing will happen with autonomous cars. Just as people expressed their luddite opinion towards textile equipment during the 1800s, people with a similar view will express their doubt towards driverless cars.
Now that you have heard both sides of the argument regarding AVs, what do you think? Would you be content on buying one?
12: Tayo Opanubi, Dee Raji, Samad Adeniran, Nicholas Tucker