Professor Ken Ueno’s latest article, “Art Beyond the AI Dataset: A Manifesto for Staying Weird,” has been published in German in Positionen 148, with the English version now available online through MF:FM.
The manifesto argues that generative AI’s ability to reproduce and recombine existing conventions should push artists not toward competition with machines, but toward forms of creativity that resist easy representation and prediction. Drawing on the history of the Japanese benshi, Édouard Glissant’s concept of opacity, and Ueno’s own embodied performance practice, the essay proposes that art’s future lies in cultivating irreducibility, embodied experience, and forms of expression that exist beyond the AI dataset.
The article also considers the implications of generative AI for artistic pedagogy. If AI models derive their generative power from historical datasets, Ueno argues, then one response for artists—and for those who teach them—is to lean into that which is not in the historical dataset: the idiosyncratic, embodied, experimental, culturally specific, and as-yet-unimagined. Rather than training students primarily to reproduce inherited conventions, pedagogy can encourage them to develop practices whose value lies precisely in their capacity to produce what cannot yet be predicted from the archive.
The German version appears in Positionen 148:
https://www.positionen.berlin/
The English version is available through MF:FM:
https://mffmonline.com/2026/08/17/art-beyond-the-ai-dataset-a-manifesto-for-staying-weird/