Patterns of Pastness. Generative AI and the Re-Imagination of History

26 November 2025 · Christoph Purschke · 2 minute read · #conversations

CuCo Lab Conversations | Roland Meyer (Zürich)

INFORMATION

Generative AI is often hailed as the technology of the future. And big tech companies are working relentlessly to make this future a reality, investing huge amounts of money and receiving enormous political support along the way. Given the unforeseeable economic, environmental, and social consequences of this vision of the future, however, it is easy to lose sight of how much AI is already changing our view of the past. Yet it is precisely their relationship to the past that distinguishes generative AI models such as Stable Diffusion or Sora from earlier media technologies: to generate supposedly new images, these models rely on images from the past. When transformed into training data for machine learning, digitized cultural archives become resources for creating alternative histories, reanimating patterns from the past, and producing nostalgic memories of events that never took place. Generative AI is thus not only structurally nostalgic, but also manifests an extractivist view of the cultural archive — everything that has ever been written, drawn, painted, photographed, filmed, composed, etc., now becomes an exploitable resource of patterns. Drawing on examples from visual culture and contemporary art, the lecture examines the political, ideological, and aesthetic implications of generative AI for our understanding of history and memory culture.

SPEAKER

COORDINATES


Close Terminal E-Mail Download GitHub Alternate GitHub Menu Check Circle Space Shuttle angle-right Warning