Konstantin Mitrokhov | The Return of the Generalist Agenda in AI research
04. Nov
6:15 pm, C40.530
The recent surge in popularity of synthetic media, such as ChatGPT, Gemini and Claude, has brought the concept of artificial general intelligence (AGI) into mainstream discourse, designating it as a threat in the new arms race. However, the generalist agenda in basic AI research is not a new phenomenon and can be traced back to the symbolic systems of the 1960s, such as the General Problem Solver. I argue for the continuity between the early work in the field and the contemporary AGI agenda, contextualising and building upon the existing critique of the latter. I discuss several episodes in the development of epistemic generality and general applicability in AI research, tracing the long-term technical and conceptual shift towards environmental rationality. Finally, I suggest that general-purpose AI systems as a genre of computation can be understood in terms of radicalisation of this environmental logic. The problem of general world representation (the frame problem) is now subsumed under the media-ecological terms, and computation becomes the frame.
Konstantin Mitrokhov is a PhD candidate and scholarship holder at Leuphana University. Having previously worked as a research associate on the DFG-funded “Browser Art” project at the Karlsruhe Institute of Technology (KIT), he is now developing his doctoral project, “Games Agents Play: On Game-like Simulations and the Possibility of Generalist Artificial Intelligence”.
Language: English
An event organized by the Research Initiative The Disruptive Condition.
Contact: Nicolas Schneider (nicolas.schneider@leuphana.de)