Files
Abstract
In this thesis, I set out a new, Kantian approach to resolving dilemmas and other conflicts of obligation for semi-autonomous machine agents such as self-driving cars. The approach begins with the modern distinction between law and ethics, and looks to a standard of justice (rather than ethics) to determine how to resolve conflicts of obligation such as in what is known as the "trolley problem." Rather than building machines that reflect one or another group's ethical preferences, efforts to build explicitly moral machine agents should focus on building rightful machines. I propose that "answer set programming," which can be understood as an efficient machine implementation of non-monotonic forms of reasoning through its answer set /stable model semantics, is a workable engineering solution for handling deontic conflicts for rightful machine agents. I critically evaluate two prior efforts in this area and demonstrate the new approach to conflicts using answer set programming.