Slides and Alternative Pasts

Larcher’s work, AI and the Art of Historical Reinterpretation (2022-2024), revolves unambiguously around ‘past reality’. Here she develops alternative pasts through a series of slides. The project aims at highlighting the contributions to history made by FLINTA, a German acronym standing for ‘women, lesbians, intersex, non-binary, trans and agender people’, and often suppressed in traditional narratives. For that purpose, Larcher used AI to effect replacements with respect to historical groups and individual portraits. 

In contrast to the series Baumeisterinnen (2018), in which the replacements of the heads of male architects with those of women architects were done by hand and remained visually evident, in AI and the Art of Historical Reinterpretation, the visual manipulation is almost invisible. The series is humorous nonetheless. With some pictures, we recall the original, while others appear familiar without our being able to identify their reference; we see a third group for the first time. The artist’s aim is to activate the sense of varied possibility in the heads of viewers, and thereby to allow them to imagine an alternative past and hence a different future. But the work also has to do with the future in another sense, namely inasmuch as Larcher reacts with this series to her investigations regarding ‘algorithmic biases’. This refers to the tendency, proceeding from AI, to produce biased outputs from distorted training data. Before the large AI companies began to work against these biases, most AI models generated, in response to such inquiries as ‘president’, only images of White men, whereas terms like ‘nurse’ were systematically rendered as female.

A quick fix in this regard consists of deliberately using so-called system prompts that remain invisible to users, in order to induce the models to diversify their output. But this changes nothing in the training data, which, since they are historical in nature, cannot readily be altered. It is here that AI and the Art of Historical Reinterpretation comes into play because Larcher imagines that the images, as a part of future datasets in training data, could contribute to less distorted results – or to an altered vision of the past. Thus, the work ultimately also touches upon a scenario in which AI-generated images are used for the training of AI models and thereby lead to undesired results.

As is shown by the discussion of these two examples, Claudia Larcher’s works explore varied points of connection to social and artistic themes. This anchoring imparts a particular interest to Claudia Larcher’s AI Art, since it is not about the impressive individual image which predominates in the flood of AI-generated images online and in the discourse concerning visual manifestations of AI, but about pictorial practices and historical relationships. The fact that certain images exist and others do not, the way they are created, and their circulation are at least as important as what they show. AI-generated images thus become part of complex experimental arrangements in whose context they can potentially be called ‘art’.

Data Set and Slide Show by CLAUDIA LARCHER

Excerpt from the text “Subtle Subversions and Beautiful System Errors” by Klaus Speidel, from the publication “Claudia Larcher – Hallucinations”, Verlag für Moderne Kunst 2025

Credits, from left to right from top to bottom:

The GM-NAA I/O system for the IBM 704 computer
G.E. Valley Jr. and three unidentified women
J. W. Forrester, N. Taylor, J. J., O’Brien, C. Corderman and N. Daggert
A. Turing
Whirlwind Computer MIT
Dr. C. G. Abbot with Harmonic printout
Woodie, MIT, 1986
Scientists and Individual Computer
Student composer N. Rowe
N. Taylor
J. W. Forrester, R. D. Ferenz et al.
E. Ferretti, left and unidentified woman
Electronic Systems Laboratory
Digitron at MIT
People Working on Computers file
Wiener and Burchard
M. Minsky, AI Lab, 1968
N. Wiener
Prof Miller seated at Computer
Scientist at MIT playing computer chess
W. Y. Tsai at MIT Hayden
R. W. Mann, MIT
P. and Marie Curie
P.O. Lindfors and F. B. Cox Jr.
I. Sutherlands Sketchpad
T. Knight and R. Greenblatt, MIT 1976
Absorption of Microwaves by Molecules of Gas, MIT 1954
Scientists at MIT

more information under:
https://www.claudialarcher.com/
https://ai-for-dignity.com/