Two of the most puzzling concepts in the history of artificial intelligence (AI), namely the AI winter and the AI effect are mutually exclusive if considered in tandem. AI winters refer to the phenomenon of loss in trust in AI systems due to underdelivery of promises, leading to further stagnation in research funding and commercial absorption. The AI effect suggests that AI’s successful applications have historically separated themselves from the AI field by the establishment of new/specialised scientific or commercial nomenclature and research cultures. How do AI scientists rebrand AI after general disillusionment in their field and how do broader computer science experts brand their research as “AI” during periods of AI hype? How does AI continue to develop in periods of “winter” in different regions’ more pleasant climates? How do periods of AI summer contribute to future periods of internet hype during their dormancy? These questions are addressed drawing from empirical research into the historical sociology of AI, a 2023 secondary analysis between technological spillages and unexpected findings for internet and HCI research during periods of intense AI hype (and vice versa, AI advancements based on periods of internet/network technologies hype), as well as a 2024 oral history project on AI at Edinburgh university and the proceedings of the EurAI Workshop on the History of AI in Europe during which, several lesser known connections have been revealed. To theorise, I am extending Pickering/Deleuze and Guattari’s notion of nomadic science previously applied to the history of mathematics and cybernetics.

Our talk investigates two early computer technologies for graphically representing molecules – the vector and the raster display – and traces their technical, material, and epistemic specificity for computational chemistry, through the nascent field of molecular graphics in the 1970s and 1980s. The main thesis is that both technologies, beyond an evolution of computer graphics from vector to raster displays, represent two modes of representing molecules with their own affordances and limitations for chemical research. Drawing on studies in the media archaeology of computer graphics and in history of science as well as primary sources, we argue that these two modes of representing molecules on the screen need to be explained through the underlying technical objects that structure them, in conjunction with the specific traditions molecular modeling stems from, the epistemic issues at stake in the involved scientific communities, the techno-scientific promises bundled with them, and the economic and industrial landsape in which they are embedded.

Frederic Wieber
Phillip Roth
Alin Olteanu

This paper centers on translation in Soviet cybernetics and AI. Focusing on cultural practices of translation and popularization as reflected in widely-read scientific and fictional texts, I interrogate practices of interpretation in relation to the professional virtue of scientific veracity as well as its didactic function in the Soviet cybernetic imaginary throughout the long Thaw. The publication of the works of Norbert Wiener, Alan Turing, and John von Neumann in Russian was not simply aimed at enabling direct access to the words and thoughts of major bourgeois thinkers concerned with automation and digital technologies: translating and popularizing cybernetics in the post-Stalinist context was about establishing new norms for public disagreement. No longer limited to the opposition of true and false positions, the debates around questions such as “Can a machine think?” that raged across a wide spectrum of Soviet media from the late 1950s to the 1980s were framed by an open-ended binate of what is meaningful or, on the contrary, meaningless. In his classic 1992 book The Human Motor: Energy, Fatigue, and the Origins of Modernity, Anson Rabinbach demonstrates how the utopian obsession with energy and fatigue shaped social thought in modern Europe. In a similar line, this project explores how human error takes on a new meaning when the ontology of information central to Western cybernetics is adopted to a Soviet version of digital modernity.

This paper explores the connected histories of whistleblowers and activists who worked in computing from the 1960s through the present day, showing how their concerns were animated by similar issues, including labor rights, antiracism, fighting against gender discrimination, and concerns regarding computing's role in the military-industrial complex. It looks at people who tried to fight the (computer's) power from within the tech industry, in order to write an alternative history of computing.