Portraits of non-existent people: AI art and (human) imagination

Abstract

The essay will address a series of portraits in which the portrayed person is non-existent. Such is the case of a number of works made through artificial intelligence (AI).
In this type of art, the machine becomes capable of elaborating on the given data in its own way, with a degree of autonomy that exceeds the human artist’s control. In the case of portraits, the AI is trained on a series of pre-existing pictures (belonging, for example, to the history of art) and becomes capable, therefore, of generating a series of new pictures, which are similar to, but also different from, the original ones: portraits of non-existent people. This case will be compared to another group of portraits Bence Nanay calls “Portraits
of People not Present”. In a paper that bears the latter title, Nanay analyzes a series of modernist portraits, in which the portrayed person is absent from the picture. This seems a contradiction because a portrait should include a representation of the person who is actually standing in front of the artist. The viewer can, however, recognize these pictures as real portraits, because mental imagery intervenes – so says Nanay – and fills the gap between what is present and what is absent. It is this idea of mental imagery that will help us demonstrate that AI art can stimulate human imagination in a new and interesting way (this will be the paper’s first objective). In fact, one of Nanay’s most important ideas seems to be that imagination plays a relevant role not only in the context of fiction but also in our everyday perception of the world. The first hypothesis of the paper will be that AI art, through
the errors and deformations of the machine, exposes our own errors and deformations in the perception of the world. AI art, therefore, brings to light the role played
by imagination in our own perception of reality: the fact that we always transform what we see and we can always see it in other ways. The second objective of the paper will be to address, in greater detail, Nanay’s idea of mental imagery. In fact, Nanay distinguishes between imagination and mental imagery. Mental imagery is different from propositional imagination because it can be voluntary but also involuntary and it is usually pre-verbal. But is verbal language really excluded from mental imagery? Or are there aspects of language that can enter this dimension? In this last case, it could be that the relationship between propositional imagination and mental imagery is not so much that of a separation, but rather a continuum between the two. This last problem will be explored through the consideration of another work of art made using AI, Klingemann’s Appropriate response. In the world of AI, as Klingemann explains, pictures and words are not two heterogeneous entities, because they are both made of pixels. What about human imagination?