Since its first formulation the Integrated Information Theory (IIT) has recently been updated to the version 4.0. Unlike the previous versions where the problem of free will was completely neglected, IIT 4.0 claims to suggest a full neuroscientific account of this oldest problem in the philosophy of mind. The aim of this opinion paper is to show that IIT’s account of free will is apparently dualist and reminiscent of the conventional free will in folk psychology, where mental constructs such as beliefs and desires are regarded as actual causes of human actions. On the other hand, these mental constructs can have high predictive power, compared to that provided by neuroscience. Thus, while rejecting ontological dualism, one can accept methodological dualism, compatible with eliminative physicalism, by virtue of its predictive power and descriptive parsimony.