The research group KIM (Artificial Intelligence and Media Philosophy) organized the lecture "Ceci n'est pas Ethics: Computational ethics and the making of machine autonomy" by Maya Indira Ganesh.
We are told that a ‘fully’ autonomous vehicle will have the capacity to make an astonishing range of decisions, from the banal to the complex, as it navigates the world. It will be able to recognise a cyclist, distinguish a dog from a ball, and follow traffic rules. It also has to figure out what to do when something unexpected happens, like a child suddenly dashing across the street chasing her dog that was chasing a cyclist. Who or what should the autonomous vehicle avoid hitting – dog, ball, cyclist or child? What if the cyclist is also a child? This scenario mirrors the conundrum generated by the thought experiment, the Trolley Problem, which has become a popular framework for the ‘ethics of autonomous driving’.
In this talk, Maya will present how the Trolley Problem, and other models from Philosophy, Economics, Applied Ethics and Computer Science, are being proposed as approaches for machines learning to compute ethics. She will discuss various typologies of computational ethics emerging therein, and the apparatuses enabling this. She argues that ethics as a computational output is both a measure of machine autonomy, and, confusingly, a measurement device as well.
Maya Indira Ganesh is a new research associate at KIM and HfG Karlsruhe in the new project “AI and the Society of the Future” that is funded by Volkswagen Stiftung.
The lecture takes place in the framework of the course "The Moral Machine: On the Automation of Ethics" by Matteo Pasquinelli. The course studies the political implications of Artificial Intelligence systems that are applied in different sectors of the economy and society to automate not only simple tasks of perception but, increasingly, also complex ethical decisions.
More information: kim.hfg-karlsruhe.de