Skip to main content
Back to KTH start page

Johan Fredrikzon

Profile picture of Johan Fredrikzon

About me

Johan Fredrikzon (M.Sc. Computer Science, Ph.D. History of Ideas) is a historian of AI and technology at the Division of History of Science, Technology and Environment, KTH Royal Institute of Technology. He writes and teaches on the flipsides of digital society: erasure, waste, forgetting, decay, death, psychosis and error. He was a visiting scholar at Stanford University (2024–2025) and a postdoctoral fellow at UC Berkeley (2022–2024), and earlier a research affiliate at Yale University.

Machines that fail. His project, The Making of Human Mistakes in the Era of Artificial Intelligence, 1940–1990 (funded by the Swedish Research Council), is a history of AI told from the perspective of error. It asks how notions of error have shaped what counts as intelligence in humans and in machines. Recent results include an article on Kenneth Colby’s “psychotic machine” (IEEE Annals of the History of Computing, forthcoming), a study of the 1972 conversation between the world's first chatbots ELIZA and PARRY (Weizenbaum Journal of the Digital Society, forthcoming), and “Rethinking Error: ‘Hallucinations’ and Epistemological Indifference” (Critical AI, 2025). He is now writing a monograph, Machines Who Fail: A History of Intelligence and Error.

New directions. Fredrikzon is developing work on “human as medium,” meaning the bodies, care and labor placed in the service of machines. Increasingly, machines are not tools that we use. Instead, humans become the media between machines, enabling them to perform. He presented it at the Dahlem Humanities Center, Freie Universität Berlin, and at CNAM in Paris in 2026. Other strands of his work look at “deadbots” and technological spiritualism, at the history of existential risk (x-risk) from AI (he designed a course on it at Stanford), and at "machine unlearning" in the traditions of the (lack of an) art of forgetting. That last strand extends the field of erasure studies he outlined with Chris Haffenden in “Towards erasure studies” (Memory, Mind & Media, 2023). A side project – Self-Developing Nation: Paper, Water, and Image in 1970s Sweden – involves using media archeology to examine the claim that in 1972, the river next to the Emsfors paper mill in Sweden was so polluted its water could be used to develop photographic film.

Books and edited volumes. His accessible history of the ideas of AI, Kiseldalen: Nedslag i den artificiella intelligensens idéhistoria, is forthcoming from Fri Tanke. He co-edits Questioning History in the Age of Artificial Intelligence (with David Bates and Julia Irwin, Fordham University Press, forthcoming), which grew out of a symposium they organized at UC Berkeley in 2024. He also co-edited Litteraturens materia (with Johan Klingborg, Glänta, 2026).

Convening. Fredrikzon initiated and co-organized the international conference series Norms in the Age of Intelligent Machines: Bodies, Knowledge, Governance with Shane Denson (Stanford) and Armen Khatchatourov (Université Gustave Eiffel). It met at Stanford in December 2025 and in Paris, in May 2026, with support from the France-Stanford Center for Interdisciplinary Studies.

In public. Fredrikzon writes and speaks regularly on AI for a Nordic audience. Recent pieces and appearances:

Earlier work. His monograph Cycles of Data: Environment, Population, Administration, and the Cultural Techniques of Early Digitalization (Mediehistoriskt arkiv, 2021) traces today’s data-driven society back to the 1960s and 70s. In that period, welfare-state agencies, insurers and research institutions were already treating data as a “raw” resource and as capital.

Selected recent publications


Private website for news and updates: www.fredrikzon.com

Profile picture of Johan Fredrikzon