A Model of Free Will for Artificial Entities
This addresses the philosophical and practical challenge of modeling free will in AI, but it is incremental as it builds on existing debates about determinism and decision-making.
The paper tackles the problem of defining free will for artificial entities by linking it to a decision mechanism where an agent first makes a predictable choice and then immediately makes an unpredictable one, replacing the intangible feeling with a concrete process.
The impression of free will is the feeling according to which our choices are neither imposed from our inside nor from outside. It is the sense we are the ultimate cause of our acts. In direct opposition with the universal determinism, the existence of free will continues to be discussed. In this paper, free will is linked to a decisional mechanism: an agent is provided with free will if having performed a predictable choice Cp, it can immediately perform another choice Cr in a random way. The intangible feeling of free will is replaced by a decision-making process including a predictable decision-making process immediately followed by an unpredictable decisional one.