← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence medium-high

Reconfiguring through Ruptures: Material Reconfigurations and Un/Making as Tangible Tactics for Queering AI-Generated Histories

Alexandra Teixeira Riggs , Noura Howell

DOI PDF Program page Page JSON

This is a compelling CHI honorable-mention contribution because it turns a familiar critique of generative AI into a specific, memorable, and materially enacted tactic. The paper’s strength is not broad validation but the clarity with which it names and demonstrates “material reconfigurations” as a situated practice for queering AI-generated histories. The evidence is exploratory and intentionally narrow, yet it aligns well with the paper’s claims and avoids pretending to offer a universal solution.

Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Review Summary

What makes this paper stand out is that it does not stop at saying generative AI misrepresents queer histories; it asks what a researcher or designer can materially do with those misrepresentations once they appear. The answer is not model tuning, prompt engineering, or abstention alone. Instead, the paper proposes a practice of generating synthetic historical images and then physically confronting them through annotation, scratching, submerging, lubricating, walking, and burying. That move gives the work a distinctive place in CHI. It is simultaneously conceptual, methodological, and artistic, but it remains grounded because the authors carefully document the process through three vignettes and explicitly frame the contribution as exploratory. The strongest reading of the paper is that “material reconfigurations” gives queer HCI and critical AI scholarship a useful new term for a recognizable but previously undernamed kind of intervention: one that acknowledges AI outputs as socially powerful representations while refusing to leave critique at the screen. The paper also productively extends archival and un/making traditions into the domain of synthetic media, arguing that the point is not to restore authenticity but to expose rupture, erasure, and false coherence. That is a meaningful contribution because it broadens what counts as method knowledge in CHI. At the same time, the paper is wisely bounded. The evidence comes from one author’s memories, three generated images, and a public exhibition, so it should not be read as proving general effectiveness across contexts. Its value lies in opening a design space and offering a rigorous example of situated critical practice. For researchers working in queer AI, archives, critical making, or first-person inquiry, this paper is likely to be generative. For readers expecting a validated toolkit or scalable intervention, it is better understood as a deliberately partial but intellectually sharp tactic that expands the repertoire of how HCI can engage generative AI critically.

What Changed

Canon before

Prior work in queer HCI, queering AI, critical making, un/making, archival critique, and arts-based inquiry had already established several ingredients this paper draws together: first-person critique of AI, material engagements with archives and histories, and refusal-oriented or subversive tactics for resisting normative systems. What was available before this paper was therefore a set of adjacent traditions rather than this exact synthesis. The paper enters a space where scholars had critiqued generative AI’s bias and where designers had used tangible practices to work through historical absence, but had not yet clearly named this specific combination of synthetic historical image generation plus embodied physical reworking as a distinct tactic for queer AI research.

Departure from common sense

The paper departs from the more obvious response to biased generative AI, which would be either to reject the tools outright or to focus on improving prompts, models, or moderation policies. Instead, it deliberately generates problematic synthetic historical images and then treats those outputs as material objects for embodied critique. The images are printed, annotated, scratched, submerged, buried, and walked with, so the intervention happens after generation and in physical space. That move is counterintuitive because it does not seek cleaner outputs or better representation from the model alone; it seeks critical understanding by exposing what generative AI cannot hold: embodied memory, queer specificity, affect, rupture, and situated historical experience.

Actual novelty

The paper’s main novelty is the explicit proposal and naming of “material reconfigurations” as a tactic for queering AI-generated histories, paired with un/making as a complementary practice. This is not just a generic arts-based reflection on AI bias. The contribution is a more specific methodological and conceptual package: generate synthetic archival images from the lead author’s queer memories, print them, annotate them by hand, materially alter them through actions such as scratching, submerging, lubricating, walking, and burying, and use those embodied manipulations to surface the ruptures, erasures, and false smoothness of AI-generated historical representation. The paper also extends Sennett’s notion of reconfiguration from already-physical archival structures to outputs of digital origin, arguing that these materially altered artifacts can acknowledge AI-enabled ruptures rather than restore an imagined authenticity.

Evidence

The evidence is appropriate for a situated CHI critical-design contribution rather than a broad empirical validation. The paper documents an autoethnographic process centered on the lead author’s memories of queer places and events in Atlanta, generating three images after up to three iterations each and then materially reworking them through annotation and physical intervention. The paper gives detailed process descriptions for model selection, prompting, printing, annotation, and the distinct material treatments used in each vignette. It also reports that the resulting video work was exhibited publicly to an audience of over 285 viewers, which supports resonance and communicative reach more than causal efficacy. Overall, the evidence supports the claim that the tactic is coherent, enacted, and reflective, while the authors explicitly avoid claiming that it is a mature or general-purpose method.

Limits

Method limits

The method is intentionally narrow and situated. It is built from one lead author’s autoethnographic memories, three generated-image vignettes, and a reflective arts-based process rather than from comparative evaluation, participant replication, or systematic outcome measurement. The paper therefore cannot establish that material reconfigurations reliably transfer across users, communities, or historical contexts. Its strongest support is for the value of the tactic as a critical and reflective practice, not for its effectiveness as a standardized intervention. The authors also explicitly limit the claim by saying these tactics are not a developed method but exploratory openings for future methodological work.

Deployment limits

This contribution does not validate a scalable deployment model. The tactics depend on printing images, engaging in embodied material manipulation, and drawing on personal memory and interpretive labor, which makes them better suited to critical design research, artistic inquiry, pedagogy, or exhibition contexts than to routine production settings. The paper also does not show organizational adoption, long-term community use, or integration into archival institutions. As a result, the work should be read as a situated critical practice for reflecting on AI-generated histories, not as a broadly deployable remediation pipeline for biased synthetic media.

Boundary conditions

The approach works best when the goal is to unsettle synthetic historical authority, foreground embodied critique, and surface what generative AI erases or smooths over. It depends on access to physical materials, willingness to engage in first-person reflective practice, and a context where subjective, queer, and materially situated interpretation is appropriate rather than bracketed out. It is especially aligned with cases involving memory, archives, identity, and refusal-oriented design. It is less suitable where the primary objective is measurable accuracy improvement, scalable content correction, or standardized evaluation, because the contribution is intentionally partial, interpretive, and grounded in lived experience rather than optimization.

Manual check: Check the pdf_url formatting against immutable metadata during merge, and verify evidence offsets against the source text because the bundle now uses multiple sections within the same full-text section. Content claims are grounded in provided spans, but publication gating should still rely on offset validation.

Position in field

The paper sits at the intersection of queer HCI, critical AI, archival critique, first-person methods, and material or arts-based design research. Within that landscape, it is best understood as a practice-level contribution that extends existing queer AI critique into a more concrete embodied tactic. It does not primarily compete with technical AI papers, benchmarks, or interface systems. Instead, it contributes a named vocabulary and a worked-through example of how researchers might materially engage AI outputs to expose bias, rupture, and historical misrepresentation. Its field position is therefore closer to critical methodology and design inquiry than to system building: it strengthens the repertoire of queer and refusal-oriented HCI by showing how synthetic images can become sites of tangible contestation rather than merely objects of textual critique.

Abstract

To critically examine the role of AI in historical representation and resist anti-LGBTQIA+ biases and erasures, we leverage un/making and propose a tactic we name material reconfigurations. We share an autoethnographic account of un/making and materially reconfiguring AI-generated images of queer histories: the lead author's memories of queer places and events. Through hand annotating, scratching, burying, submerging, and walking with physical images, they un/make and reconfigure, highlighting embodied aspects of archival records unparsable by generative AI. We propose that un/making and materially reconfiguring synthetic archival images can resist generative AI's increasingly hegemonic role in misrepresenting historical data and erasing queer identities. We contribute reflections on un/making and material reconfigurations as tangible tactics for queering AI, attuning to queer temporalities to unsettle AI-generated histories, using embodied, autoethnographic practices as critical strategies, and working through tensions of use and refusal in Queer AI research.