Ratoff2022Self-DrivingCars

William Ratoff, "ESDiT-shared - OLD 2021-11-03/esditHuman/self-driving cars and the Right to Drive"

Bibliographic info

Ratoff, W. (2022). Self-driving Cars and the Right to Drive. Philosophy & Technology, 35(3), 1-15. (APA)

Commentary

⇒ In general, what about this text makes it particularly interesting or thought-provoking? What do you see as weaknesses?
This text addresses the question regarding whether humans should be allowed to drive cars if there are self-driving alternatives that are less dangerous. What makes this text thought-provoking is that currently we are seeing that AI applications are becoming better than humans at specific tasks. This also holds for driving, previous research has shown that self-driving cars tend to cause less accidents (and thus fatalities) than human-piloted cars. In the field of digital ethics, this has brought up the question whether humans should still be allowed to drive cars if there is a safer alternative. The main argument contra seems to be that by allowing human-piloted cars on the road, other individuals are exposed to unnecessary risks. In our current legislative system this is not allowed and therefore one could argue that only AI-driven cars should be allowed on the road. However, the author of this paper shifts the discussion towards a more nuanced question. One could argue that on a daily basis, individuals undertake actions that expose other individuals to unnecessary risks. However, due to the fact that these risks are so small, our legislative system does not forbid these actions. If we look at the problem through this scope, the question becomes whether human-piloted cars impose a significant unnecessary risk to others instead of a risk in general. I think that it is important to evaluate every instance of AI replacing human tasks separately as more factors are important to consider than just the improvement in efficiency. For this specific case, although AI-piloted cars are less harmful than human-piloted cars, the author suggests that the difference is too small to take away the freedom and choice that humans have regarding the driving of cars.

To conclude, I think that the author points out an important nuance in the debate concerned with the adoption of self-driving cars. However, the final argument he makes leading him to conclude that human-piloted cars are no significant unnecessary risk to other individuals is not very convincing. I attribute this to the fact that the argument heavily relies on the projected similarity between driving a car and 'going for an enjoyable spin without further purpose'. To make this argument more convincing, it would have been better to focus on a more theoretical argument instead of a single normative comparison between two activities.

Excerpts & Key Quotes

Defining the harm principle

"In essence, the Harm Principle has it that people should be free to act however they please so long as their actions cause no harm to anybody else. The state — or anybody for that matter — has no business in regulating any activity that yields no (third-party) victims."
"It articulates a limit on the permissible reach of the law: the state is warranted in legislating against some activity only if that activity poses a harm to others."

Comment:

This passage explains Mill's Harm principle and also forms the foundation of the argument made by the author. It should be noted that the Harm principle gains most of its meaning once agreement has been obtained regarding the definition of 'harm to anybody else'. Harm can have many definitions and if a very light definition is chosen, this can result in the state interfering with every activity undertaken by individuals. The author therefore argues that it is more desirable to choose a more strict definition of harm (one that aligns with the infringement of an individual's rights as defined by our legal system). The principle revolves around the notion that the state should not interfere with its citizens behavior if that behavior does not substantially affect others in a negative way. In the context of this paper, this implies that the state should or should not be allowed to mandate the use of self-driving cars.

The argument for human-piloted cars

"(1) The state is warranted in legislating against some activity A only if that activity A violates the rights of third-parties.
(2) A driver S who chooses to drive herself on public roads rather than using self- driving technology does not violate the rights of any third-parties in so acting.
(3) Therefore, the state is not warranted in mandating use of self-driving car technology on public roads."

Comment:

This passage contains the argument that the author proposes. The author argues that the state should not be able to legally mandate the use of AI-driven vehicles in the public space. The first premise stated by the author is a simple repetition of the Harm principle as presented by Mill. Secondly, the author argues that human-piloted cars do not pose a violation to the rights of third-parties (on the road). This premise is based on a comparison given by the author in which he suggests a similarity between driving a car and going for a 'spin' with a care for leisure-based purposes. The author argues that because the latter of the two does not pose a significant risk to third-parties nor does the first. Lastly, the author concludes that the state should not mandate the use of self-driving cars, this follows from the first two premises.

Where to draw the line regarding the exposure of 3rd parties to unnecessary risks

third-party harm

"However, I’m skeptical of this line of thought. On what grounds? In essence, I don’t think that a driver who chooses to drive herself, over relying on self-driving car technology, is violating other road-users right not to be subject to unnecessary risks. Why? Well, although it is indisputable that by choosing to drive herself she is creating extra unnecessary risk for third-party road-users, it’s far from clear that this behavior violates the right of these third-parties not to be subject to unnecessary risks. After all, it is intuitively obvious that not any imposition of unnecessary risk violates this right".

Comment:

This passage has been chosen because it elaborates further on the second premise stated in the previous passage above. The core of the argument made by the author is that not all instances of creating unnecessary risks for third parties are of such magnitude that they violate the right to exposure to unnecessary risks. This intuitively makes sense, for instance, if I ride my bike through the city, I am to some degree exposing others to the risk of collision. However, it would be nonsensical to forbid individuals to ride their bikes to the city. This is because the exposure to risk for third-parties is so small that the act of riding a bike does not violate the legal definition of exposing others to unnecessary risks. In some sense, one could argue that the risk of certain activities is negligible. The final stance of the author seems to be that this is also the case for humans driving cars. A statistic that he provides is that 1 fatality occurs per ~1 million miles driven, the author concludes from this that human-piloted risks form a very small, possibly negligible risk to others.