Prerequisites for a Discussion on the ‘Free Will’ of Artificial Intelligence: Buddhist and Science-based views of “Ex Machina”

This paper discusses the prerequisites for deliberations on the free will problem of artificial intelligence, with reference to Alex Garland’s 2015 film “Ex Machina.” The implications of Buddhist non-self theory are examined against the common idea that free will originates from the self. From a viewpoint common to Buddhism and neuroscience, possible interpretations of the film are suggested based on the results of recent robotics experiments and technological feasibility. The Buddhist doctrine of no-self does not allow an agent à la ātman. The concept of ‘self as a useful fiction’ facilitates the notion of ‘action without an agent.’ This concept is also useful for explaining the integrated internal representation of ‘I’ or the domain-general thinking ability. In the film, Ava behaves comfortingly in human-suitable spaces in a human-like way. That means that Ava’s self-model has achieved same level of usefulness as the human self-model. It seems no longer possible to argue that artificial intelligence cannot have a ‘mind’ solely by the fact that it is constructed differently from humans. The boundaries between living things and non-living things have become blurred and many-layered. However, there are clear differences in the cognitive structure of humans and artificial intelligence systems. Current artificial intelligence technology can read, analyze, and generate emotional expressions. The emotional experience and self-reflective consciousness of artificial intelligence remain hard problems from both practical and theoretical standpoints. In discussions of the free will of artificial intelligence, it is not necessary to presuppose the idea of the self like ātman; instead, the experience of emotion and the generation of consciousness are conditions that must be considered before exploring the issue of free will.