Generating Queer Histories

Prompting Critical Reflections on Generative AI in the Archives

Activities: Research through Design, Autoethnography, Video Production

Year: 2025-26

Year: 2019

"Generating Queer Histories" critically examines the role of AI in historical representation and reflects on how AI might be un/made or reconfigured to reconceptualize archival ephemera from personal queer histories. In this project, I use Gen AI to create images of archival records–ephemera, places, or memories evocative of queer history. I then annotate, question, collage, un-make, and reconfigure these visualizations by hand. By annotating and unmaking, I materially reflect on Gen AI’s hegemonic role in representing historical data. Further, by bringing this process into tangible, material realities, I highlight the intimate, emotional, and embodied impressions that accompany archival records but are often erased by or difficult to parse in algorithmically defined databases. Similarly, by using photographic collages, or ephemera, I turn attention back to queer identities that have been “locked out of official histories” or material realities, to critically reflect on what “counts” as part of the historical record. 

Publications

Tangible tactics

resisting algorithmic erasure

In this project, I critically examine how AI mediates historical representations and propose tangible tactics—un/making and material reconfigurations—for resisting anti-LGBTQIA+ algorithmic erasure by highlighting embodied aspects of archival records unparsable by generative AI. 

Un/making engages a spectrum of making and unmaking to invite reflections on queer identities and histories. Material reconfigurations are embodied processes that reinterpret synthetic archival materials to acknowledge AI-enabled ruptures and resist the misrepresentations of an imagined past.

Prompts_3up_noborder

In the project, I used GenAI to create images of archival records–ephemera, places, or my own memories evocative of queer history, specifically in Atlanta. The first step in my process was to select models and prompt to generate images.

Vignettes1-3_annotations

I then handwrote annotations and drew onto the printed synthetic images to express critical perspectives. For instance, in Vignette 1 (left), my annotations called out the disconnect between my memory of a raucous, diverse queer event, and the eerie homogeneity of the generated image. In Vignette 2 (center), my annotations focused on the generic sexual wellness items generated in the image, along with my surprise at my ability to generate these forms through workarounds. In Vignette 3 (right), I focused on the lack of queer representation beyond White cis men, the lack of location specificity, and the stereotypical scenery choices in the now shuttered queer bar near my apartment.

Un/Making & Materially Reconfiguring

expressing critical perspectives

After annotating, I un/made and materially reconfigured the generated images through a series of physical processes (shown below). Doing so, I tangibly reflected on generative AI’s normalizing role in representing historical data and highlighted the intimate and embodied sensibilities that accompany archival records but are either erased or smoothed over. This work suggests how tangible practices of un/making and material reconfigurations can work against the influence of algorithmically mediated archives, towards critical reflections on what “counts” as part of the historical record 

Vignette 1_wide

Vignette 1: Queer Lube Wrestling 
Un/making the image by submerging it in water, to create a materially reconfigured piece, through the accumulation of dirt, dust, and water over time. Watching the photograph fade and decay served as a reminder of GenAI’s reliance on natural materials (e.g. water, land, carbon), and the process subverted a dynamic where generative AI extracts resources, conceptually imagining how natural processes might instead erase the image from the page.

Vignette 2: A Queer-Centered, Disability Inclusive Sexual Wellness Store
Un/making the image by lubricating and scratching it to create a materially reconfigured piece that shows lube, scratches, ruptures, and altered synthetic image. Stripping the photographic image from the paper in this way reflected my frustrations at ChatGPT’s content moderation and physically enacted the digital erasure I sought to critique. This violent scratching symbolized the epistemic violence of erasure and misrepresentation and inserted cheeky material subversions into content moderation policies.

Vignette 2_wide
Vignette 3_wide

Vignette 3: A Permanently Closed Bar
Un/making by walking with and burying the image to create a materially reconfigured piece through the enactment of this process and the entangling of the paper with the earth. Through this enactment, I urged reflection on the dissonance between the closed bar and the straight-washed synthetic portrayal of the scene.

Conclusion

provocations for queering AI

Through this project, I share how un/making and my proposed material reconfigurations can serve as tangible tactics for queering AI—specifically for critiquing biases and erasures in generated archival images.

These tactics surface the following provocations for queering AI:

  1. Expressing embodied dimensions of records pushes against the primacy of algorithmically defined archives
  2. Synthetic pasts reverberate into futures that are increasingly defined by the absence of marginalized people—if we cannot exist in generated pasts, our power to affect present and future realities diminishes
  3. Un/making and material reconfigurations can be used as techniques to reassert the tangible and affective dimensions that cannot be conveyed via generative means, giving design and HCI researchers agency in expressing embodied unease

 

Allie Teixeira Riggs

Allie Teixeira Riggs

Allie Teixeira Riggs

Design Researcher and Product Designer

Design Researcher and Product Designer

Design Researcher and Product Designer

Design Researcher and Product Designer

Email   ariggs00@gmail.com

LinkedIn  atriggs