The Pros And Cons Of Autonomous Vehicles

Liability

Autonomous vehicles can raise a lot of important ethical questions. According to google (“Ethics”, z.d.), one of them is liability. Who can be held responsible if an autonomous vehicle crashes? Is it the driver? Google? The programmer? Not only google but also by other companies like BMW it is considered to be a serious dilemma. A recent study (Marchant & Lindor, 2012) showed that holding the car manufacturers responsible for any crash caused by an autonomous vehicle would probably be the best solution. This because they are, after all ultimately responsible for the final product. In addition, it also has its advantages. Marchant & Lindor (2012) stated that “the liability burden on the manufacturer may be prohibitive of further development”.

There are even four reasons for a collision involving an autonomous vehicle, shown in a recent study (Fraichard & Jr, 2012), namely hardware failures, software bugs, perceptual errors, and reasoning errors. Hardware failures are some-what predictable and often gradual, while software failures are unexpected and sudden. This can cause dangerous situations when driving at high speeds. A perceptual error can emerge through the misclassifying of an object on the road. In addition, Fraichard & Jr. (2012) state that “if a pedestrian is correctly classified, an automated vehicle would need some way to perceive her intent, for example whether she is about to step into the road or is merely standing on the sidewalk. A mistake in this calculation could lead to a crash, especially considering the close proximity and high speed differentials on roadways”. Therefore can be considered that the system is somewhat imperfect, certainly the first generation of autonomous vehicles. This can lead to ethical decision making. In the study of Goodall (2014) they quoted the theory of the trolley-problem in which a person must decide whether to switch the path of a trolley onto a track that will kill one person in order to spare five passengers. It’s a hypothetical example that may only be needed in extremely rare circumstances. Google on the other hand came up with some interesting moral questions according to Goodall (2014). What

(2014), autonomous vehicles need robust security protocols. This will be a very challenging task for the several car manufacturers and there communication platforms. Anderson et al. (2014) noted “the need for proper protocol layers and the ability to make the communications solution scalable across hundreds of thousands- possibly even millions- of autonomous vehicles”. It’s needless to say that these communication platforms are very vulnerable to different types of security attacks. In addition the requirements may be a potential bottleneck for mass deployment. It namely affects the privacy of the owners of the autonomous vehicles. There is a lot of data that’s being collected by the in-vehicle communication platform about as well the vehicle, as the driver inside. Anderson et al. (2014) state that insurance companies for example would be very interested in individual driving habits and retailers would be very interested in attracting motorists to their locations. In addition law enforcement agencies also would have high interest in the data collected by these in-vehicle communication platforms. Yet it’s still not sure who will own the data, collected by the autonomous vehicle as they move, gather and transmit information. A possible solution is the anonymization of the vehicle according to Anderson et al (2014). There are already several manufacturers of autonomous vehicles working on the aggregation of data, so it would not reveal