Title

Russell and Floridi on Artificial Intelligence and Agency

Abstract

Are (some) AI technologies agents and if so, in what sense(s)? And why consider this topic? One interesting similarity between Stuart Russell’s discussion of AI in his book Human Compatible (2019) and Luciano Floridi’s discussion of AI in his book Ethics of Artificial Intelligence (2023) is that both of them define artificial intelligence in terms of agency. In my presentation, I will first motivate why it is interesting to reflect on whether at least some AI technologies are agents and then compare and contrast the agency-based definitions of AI in the Russell and Floridi’s respective books. I will argue that their views can be combined in interesting ways but that it is best to define AI in a way that leaves it open whether AI technologies are agents in any significant sense.

About Sven

Sven Nyholm is Professor of Ethics of Artificial Intelligence at the LMU.