Abstract

The outputs of generative artificial intelligence (generative AI) are often called “synthetic” to imply that they are not natural but artificial. Against that use of the term, this article focuses on a different denotation of synthesis, stressing the unifying and compositional aspects of anything synthetic. The case of large language models (LLMs) is used as an example to address synthesis philosophically alongside notions of representation in contemporary computational systems. It is argued that synthesis in generative AI should be understood as a search for unity that is fundamental to the making of a representational reality. This representational reality, internal to computation, is a stable (if imperfect) whole, a togetherness of distributed representations. The article thus demonstrates that developing the philosophical concept of synthesis to investigate today’s generative AI involves examining how structuring occurs in LLMs and studying the kinds of forms that structuring results in.