If we consider the aesthetic productions of artificial intelligence (AI) to be artifice, and not art (defined here as subjective self-expression), what are some reasons why?
1/ There isn’t a subject that expresses itself. AI learns by algorithmic mimicry and exceeding the limits of that mimicry - the AI’s processing horizon - won’t yield subjective self-awareness, not, at least, in the way modern philosophers of consciousness define it (i.e. “There is a way it is like to be an AI”). Animals (including humans) and AI can be considered algorithms, but for a human being, the comparison is metaphorical. There isn’t any persuasive evidence that AI will ever understand or relate to itself metaphorically, through a mediating language of symbolic self-reference. So far, the ability to make metaphors is a human trait that can only pass to other humans.
2/ It is true that AI has been able to generate novel productions. But on finer scales this is still algorithmic mimicry, a pastiche created from the available repository of human cultural artifacts. AI does not exceed the bounds of that repository, even if it finds the bottom, because it has no imagination, which is a uniquely human endowment enabling spontaneous and ineffable expressions of creativity in the arts and sciences, and in life generally.
3/ An artist refines their art - their individual expression - because it’s part of their agency in the world. Criticism, self-criticism and editing are essential to human creativity, even if they are denied during a spontaneous creative act. AI can generate more efficient versions of its algorithmic mimicry, true, but it still can’t exceed the bounds of its processing horizon, which negates uniquely subjective self-expression and criticism thereof. The ability to employ subjective critique to evaluate and improve on existing productions remains, for the present, the purview of human beings, although the efficiency of AI in implementing human-directed aesthetic choices is certainly superior overall.
