(Un)stable diffusions

Generative AI is a uniquely public technology. The large language models behind ChatGPT and other tools that generate text and images is a major develop in publicity as much as technology. Without public data and public participation, these large models could not be trained. Without the attention, hype, and hope around these technologies, the big AI firms probably could not afford the computational costs to train these models. Our special issue questions how Critical AI Studies can attend to the publics, publicities, and publicizations of generative AI. We situate AI’s publicity as mode of publicity – hype, scandals, silences, and inevitability – as well as a mode of participation seen in the grown importance of technology demonstrations. Within this situation our contributions offer four different research paths: (1) situating the legacy media as an enduring process of legitimation; (2) looking at the ways that AI has a private life in public; (3) questioning the post-democratic future of public participation; and, (4) developing new prototypes of public participation through research creation.