One objection related to the “Platonic Representation Hypothesis”, which essentially identifies below Artificial Intelligence models a set of “universal ideas” common to spoken languages and images recognized by humans, derives from the observation that language can express a concept or feeling that many images cannot. It should also be noted that in many cases language cannot replace an image in describing a visual concept.
Hence the need to develop new systems of both visual and textual representation.
Here we are helped, for example, by reference to Chinese ideographic writing, which manages to express in essentially visual form the entire dictionary of a language, including abstract concepts. But Chinese is not a universal language; it is not recognizable by everyone and expresses a specific linguistic and visual culture.
In a future scenario in which in the sphere of Artificial Intelligence language and vision align, converging in a universal “world of ideas,” visual arts could find a new historical role: creating new forms of universal “ideo-graphies,” common writings shared with humans and machines.
Italiano