Self-Driving Ethical Dilemmas

Who is Responsible?

With self-driving cars on the horizon for the average consumer, an ethical dilemma is made apparent. Who is to blame in the advent of an accident involving a car that drives itself? There are many situations in which the car could make a “wrong” judgement call based off of its internal decision models. The problem occurs when the decision the car makes differs from what the average person would consider a good moral choice. A current issue relates to who is responsible for an assisted driving accident, when the car attempts to save the driver or a pedestrian, and the other issue involves self-driving cars and the responsibility of computer model decisions.

The article by Patrick Lin addresses the issues with assisted-driving cars attempting to make ethical decisions in the

event of an accident. He proposes a bizarre theory which involves a system that the consumer may adjust to draft out a ethical decision model that matches the driver’s ethics. The solution this system solves is how to create a perfect ethical decision model for an assisted driving car that will align with the consumer’s ethics in the attempt to limit the liability of the car manufacturer. The biggest issue with this model is that there is no way to perfectly align the model with a driver’s every ethical decision. So there still exists a bias created by the car manufacturer that would influence the car’s decisions.

One of the biggest problems that car manufacturers are trying to solve is how to limit their responsibility as some of the decisions that car could make have the possibility to be counter to what they originally predicted in their software.

For instance, some new software and decision models that are currently being developed in advanced computer science, are neural nets. These neural nets imitate the human brain by tweaking statistical averages for nodes to compute a decision based off of a set of input data. This input data would be things like where the car is, how many cars are around, speed limit, etc. The biggest usage with the neural net is that it can learn and develop its own decisions based off a set of input data that identifies a basic set of decisions. With these neural nets at the heart of the decision models, the car manufacturer sometimes can not predict what the neural net will decide as the neural net has essentially learned a decision model and drafts its own conclusions for a incident. So the car manufacturer has the possibility to no longer be involved in the decision making model and can no longer be responsible for the car’s “supposed” immoral

decision.

There is also the issue that the driver can no longer be responsible for the actions of the car when cars become fully autonomous. The people who purchase cars will no longer be drivers but simply passengers with destinations. So as stated with the car manufacturers having less of an impact in the decision making model of the car and the driver transitioning into simply passengers, who is responsible?