Charles Rathkopf: Does AI vindicate functionalism?

The upcoming session of the “Philosophy of Cognitive Science” seminar is scheduled for November 9th at 10:30 AM (Warsaw, CET). The seminar meeting will be hybrid: in person in Pałac Staszica (ul. Nowy Świat 72, room 161) or via Google Meet. Our distinguished guest for this session will be Charles Rathkopf. We will delve into the intriguing topic of his paper draft, titled “Does AI vindicate functionalism.”

Abstract: Experiments in the nascent field of NeuroAI suggest that some deep neural networks are capable of executing some of the same operations as the human brain. Consequently, these experiments appear to offer unprecedentedly direct evidence for a liberal version of the functionalist position in the philosophy of mind, according to which mental properties can be realized in non-biological systems. I scrutinize this hypothesis by subjecting it to a stringent test from the literature on multiple realization. It turns out that the empirical assessment of this kind of multiple realization yields indeterminate results, because it is not clear whether the mental kind in question is genuine. This indeterminacy suggests that the claim that a computer carries out the same mental operation as some component of the human brain has less empirical content than functionalists have tended to presume. I conclude by describing how this result extends and sharpens a family of older “triviality arguments” against functionalism.

You can also join us online via Google Meet. Just ask Marcin Miłkowski at

Print Friendly, PDF & Email


Log In

Create an account
Europejski Sondaż Społeczny | European Social SurveyEuropejski Sondaż Społeczny | European Social Survey