Spring 2026: “Slop” and “Nostalgia”

The conference will be held on Zoom, and links will be sent to all registered presenters and attendees the day before the conference.

To register to attend the conference, please complete the form here.

Welcome and Introduction

DLC+ Banner reading slop and nostalgia, date 14 May 2026, as part of the series of Current Keywords in Digital Literary Culture Mini-Conference

 

Conference Start and Welcome:

10:30 AM EST / 16:30 CET / 9:30 CST/7:30 PST

Slop

In recent years, the term ‘slop’ has emerged as a keyword in public discourse, condensing a range of anxieties around the widespread integration of generative AI technologies into everyday life. In the context of social media platforms, the term slop is used to describe the proliferating stream of low-effort, low-quality content that now populates algorithmic feeds, sometimes characterised by a flattened or ‘homogenized aesthetic’ (Madsen and Puyt 2025). Slop is also frequently framed as a symptom of the broader ‘enshittification’ of the internet: a moment when widely available generative AI systems and platform incentives combine to produce a rapid degradation of both content and user experience (Doctorow, 2025).

This paper argues that slop should be understood less as a category with clear aesthetic or generic boundaries, and more as an affective keyword through which users articulate the lived experiences of the ‘generative AI revolution’ (Leslie and Meng, 2024). Bringing Raymond Williams’s account of keywords as sites where emergent cultural meanings crystallise into dialogue with Lauren Berlant’s reflections on affective atmospheres, the paper reads the discourse surrounding slop as a process of attempting to name and negotiate the lived experience of AI-saturated platform media environments. Drawing on examples of AI slop from content creators on TikTok and Instagram, user commentary on platforms and user forums, and online journalism, the paper examines the affective tones and emotional valences that circulate around slop as both content and discourse in platform media environments. By focusing on the affective scenes that recur around slop discourse—from disgust and doom to mocking derision and disbelief—I consider how slop condenses a range of everyday encounters with algorithmic feeds, generative technologies, and automated cultural production into a shared affective vocabulary for making sense of life in the age of ubiquitous AI.

Kommers et al’s recent article “Why Slop Matters” (*ACM AI Letters* Jan 2026) cites Clement Greenberg’s 1939 essay, “Avant-Garde and Kitsch,” as the basis for a historical analogy for slop’s place in the cultural field: it’s a safe bet that slop will make a historically familiar trek from a reviled, emergent cultural form toward an established, versatile, and dominant one. Don’t be wrong about slop like Greenberg on Rockwell, they argue. This paper aims to recalibrate that historical frame of reference, arguing that we’re not there yet: slop isn’t the art in Greenberg’s 1939, or even yet the film in Walter Benjamin’s 1936, but rather the instant photography and film in the 1890s.

The analogy offers two quick directions. First, it seems reasonable to consider a medium’s novelty phase as a phenomenon well worth studying: the film styles of Méliès, Lumière, Edison, and the broader cinema of attractions explored techniques and subjects that ultimately differed substantially from the film editing grammar, narrative style, and storytelling strategies of even the 1910s. I consider as a brief contemporary example *Eno* (2025-6), a generative documentary whose approach cuts against that of most GenAI content. It’s generative in its editing of historic footage, rather than images. Expressive GenAI is in its infancy, before aesthetic evaluation even quite makes sense. Drop “slop.”

The second, shorter element of a historical analogy to the 1890s takes Warren and Brandeis’s “The Right to Privacy” (1890) as a consideration of the practical dimensions of GenAI content: instant photography created new opportunities for abuse, harassment, and exploitation. Likewise, GenAI images create new ways for information about us to circulate without our consent. With new media tech, the artistic potentials of a form must be situated in a broader ecology of practical use and abuse. (Fuck Grok.)

Our paper is an exploratory work debating a research question informed by media archaeology and archive studies: can training or tuning generative image systems with “excavated” media memories reopen the model’s visual possibility space and partially counter slopification? Excavated images include screenshots of early web pages retrieved via the Wayback Machine; captures of obsolete graphical user interfaces, iconographies, and typographic defaults; and also digitized broadcast remnants such as recordings of television stations circulating on YouTube. These materials usually retain different layered inscriptions— absent ou partially rendered images, VHS texturized recordings, edited contemporary information —But without explicit curatorial and model-training strategies, irregularity may be smoothed away as noise, or stylized into a kitsch “web-vintage” preset that reproduces homogenization by other means leading, potentially, to a sloppifcation of these “genre” of images. We therefore outline a preliminary speculative research design that treats datasets as operational archives: (1) stratified corpus construction by time, platform, format, and provenance, preserving contextual metadata; (2) annotation schemes that distinguish categories of irregularity (interface traces, transmission noise, platform compression, and archival capture artifacts); (3) sampling and weighting strategies that keep such traces legible to the model; and (4) ablation studies comparing models trained with and without stratification, labels, and weighting. To evaluate whether excavated training materially alters generative behavior towards sloppificatoni, we propose a mixed-method assessment combining quantitative diversity measures with qualitative, interpretive readings grounded in media archaeology. Task-oriented prompts focused on “images about media”—screens, archives, devices, interfaces, and breakdowns— could test whether models can render historically situated mediation rather than defaulting to generic polish. The goal is not novelty for its own sake, but the capacity to produce temporally and materially specific images of media infrastructures. As a starting problematization rather than a turnkey solution, the article reframes generative image quality as a politics of memory: asking which pasts are made legible when archives become training data, and which are filtered out as “low quality” or becoming “slop”.

In the context of a contemporary digital popular culture that prompts “most people to construct a public identity,” the new media persona requires “technologies of computation and mediation and their interfaces that function to automate, produce and filter communication” (Marshall 7). I will interrogate how “lore,” whether used in direct address to a camera and its imagined audience while self-narrating a video titled “Part 6 of ???” or to reference a body of knowledge or past events associated with a figure, whether friend, fictional or famous, reveals attempts to “construct a coherent identity” from fragments whose “units constitute self-representation” (Gazi 10). Conceptually, I will situate lore as a device to perform a sort of “self-serialization.” Marshall called seriality “elemental to online culture” and structured its “constitution of person and persona” and critically broadened it “to identify anything that is in sequence or linked or joined” (Marshall 4). From the printing press to digital cultures, seriality remained central to literary texts, genres, and “photography, cinema, and the television format” (Sielke 44). Still, more than “a production and publication logic,” serialization is “a narrative form that performs important cultural work and has profound ideological and aesthetic consequences” (Allen and Van den Berg 3).

From tech venture capitalist Bryan Johnson using his biomarkers to “anti-age” himself to the transhumanism movement, Silicon Valley’s use of AI to optimize the body is dependent on automating – thus maximizing – surveillance, based on the assumption that more data means more knowledge. In the generative AI technology from these Silicon Valley companies such as OpenAI and Midjourney, the diffusion models are trained by scraping millions of images and learning statistical patterns of visual probability, encoding what bodies most frequently look like across the data landscape. Here, the relationship between the visual and computation become apparent; the generated body is a product of the desire to optimize not only the physical body, but also the data extracted from it. When this optimization process encounters bodies at the edges of its training distribution — unusual poses, multiple overlapping figures, extreme proportions — it produces the characteristic failures of slop: extra limbs, melting skin, nonsensical spatial relations that reveal the statistical seams of the model. We argue that AI slop is the failure of optimizing the biological body in its creation of bodies that appear and move nonsensically, illogically, and beyond bodily limits – thus suboptimally. This failure of optimizing presents a crucial opportunity to reject conforming to the standardization of bodies. Our presentation will establish how the ideal body becomes a produced image by tracing the natural body in Leni Riefenstahl’s Olympia (1938) to the artificial aesthetics of body modification that appear in generated bodies. In doing so, we will demonstrate how failure of optimization in early and current AI slop rejects the desire to maximize the body, thereby the data extracted as surveillance capitalism amplifies the pervasiveness of data collection. Lastly, we will present an original artwork that encapsulates how this failure is becoming an intentional aesthetic choice to “gamify” the attention economy.

Break

Approximately 12:00 – 12:15 PM EST/ 18:00 – 18:15 CET / 11:00 – 11:15 CST/ 9:00 – 9:15 AM PST

Nostalgia

In the _The Dictionary of Obscure Sorrows_ (2012), John Koenig coined “anemoia” to name “nostalgia for a time or place one has never known.” Teaching at Skeuomorph Press (https://skeuomorph.ischool.illinois.edu/), I encounter anemoia frequently, as students enthuse over movable type, printing presses, typewriters, and vintage computers, all obsolete long before they were born. These responses have become only more forceful during the past few years, as students increasingly situate their work with our “dead media” as responses—whether humorous, serious, or sanctimonious—to the emergence of computational—and especially AI—systems in their educational and professional lives. Our most frequent patrons this academic year are members of the Illinois Luddite Society, who print broadsides declaring “RESIST SHITTY TECH!” using bold wood type and stark black ink. There is irony, of course, in using this particular technology to make such declarations: letterpress is only artisanal by contrast to digital technologies, but was in its heyday primarily a technology of industrialization. This paper attempts neither to lionize nor mock these complex and conflicting media impulses, but to unpack what Alan Liu calls “the _déjá vu_ haunting of new by old media.” How do nostalgia, anemoia, and one of their products, skeuomorphism, shape our own moment of media transition, from these student responses to corporate branding, such as OpenAI naming their coding platform Codex? The notion of a skeuomorph as flatly nostalgic or anemoiaic assumes a directionality of media change: either away from a glorious past we should hold onto, or backward from a future we should aspire toward. By contrast, this paper proposes a theory of skeuomorphism as embedded cultural critique, whether deliberate or inadvertent. The skeuomorph speaks to which cultural values, aesthetic preferences, or expressive possibilities users see as vital to retain through new media shift.

In my paper, I will introduce the contemporary periodical genre of the mindstyle magazine as a postdigital metatext of digital media. Coined as a counterpart to ‘lifestyle,’ mindstyle publications such as Flow, Breathe, Happinez, and Oh (since the 2010s) primarily target women readers, reconfiguring the conventions of the women’s magazine by rejecting its aspirational consumerism and body-centric narratives. Instead, they tap into the affect and emotions of the ‘happiness industry’ (Sara Ahmed 2010) and the ‘digital backlash’ (Kristoffer Albris et al. 2024), focusing on well-being, interiority, and creativity: Text, images, and design often hinge on the opposition of online and offline experiences, urging readers to engage with, e.g., book reading and nature as offering more ‘valuable’ and restorative experiences.

Casting themselves as an analogue escape, the magazines’ design too must be understood against the foil of digital media: uncoated stock and papery extras like postcards, posters, and ephemera foreground tactility. Ornaments, design, and images evoke a predigital, peacefully analogue world, with illustration styles leaning towards hand-drawn and child-like imperfection, suggesting the aura of the handmade and authentic object as well as an unspecified ‘past’ period.

Flow magazine’s annual spin-off publication, The Flow Book for Paper Lovers, takes this a step further, consisting solely of paper objects that can be detached and used for craft projects. Images and ornaments are cute, cosily domestic, and feminine, depicting scenes and motives conspicuously devoid of screens. Mindstyle magazines, I argue, reveal contemporary fantasies and fears about digital culture, and the binary between digital media and analogue unfolding a magazinal version of what A.R.E. Taylor (2006) has termed the ‘analogue idyll.’

In the design of their games, many mascot horror developers have mentioned the role nostalgia played in the creation of their game worlds. The trappings of childhood, in particular, appealed to the developers as a shared experience their prospective players could relate to, which the game can then render uncanny. With the majority of mascot horror developers born within the Millennial and Gen Z generations of the global north, however, this “shared experience” is a specific one—that of the highly consumerised, pre-Internet childhood of the late 20th century.

Previous research on mascot horror—and the Five Nights at Freddy’s franchise, in particular—has positioned these games as a part of the digital gothic genre, due to their emphasis on media obsolescence and the instability of memory this leads to in the digital age. But little attention has been paid to the ways mascot horror performs what Svetlana Boym describes as reflective nostalgia; a form of nostalgia that questions the supposed truths of the past.

This presentation will interrogate the ways in which mascot horror calls into doubt the warmth of feeling for analogue childhood by revealing the cold reality of fast capitalism and consumerism, which have led to the contemporary, commodified self. This exploration will be performed via a close reading of the monsters, settings, and narratives of the Five Nights at Freddy’s and Poppy Playtime transmedia franchises. Ultimately, this will expose how mascot horror instrumentalises the cultural anxieties of late capitalist society to achieve its own commercial success.

A photograph is a ‘temporal hallucination’, Roland Barthes tells us in Camera Lucida (1980) of the product of photography, a medium which fixates in time a moment that can never be found again. Photographs are at the heart of Agrippa (a book of the dead), a 1992 hybrid – part-analogue, part-digital – artist’s book meant to self-destruct in the act of reading, thus fixating in time the moment where a reader unfolds Agrippa’s story. The joint work of cult science-fiction author William Gibson, visual artist Dennis Ashbaugh and publisher Kevin Begos Jr, Agrippa presents itself in the codex form, enclosing images by Ashbaugh made to fade over time, and as one leaf through the pages. They are visual responses to an early electronic poem by Gibson, based on his discovery of a family photo-album, and enclosed in a diskette programmed to be only read once. In Mechanisms (2008), Matthew Kirschenbaum defined Gibson’s poem, ‘Agrippa’, as ‘closest to a literal embodying of the peculiar nature of electronic objects [..]: their remarkable staying power and their fugitive abandon.’ This talk will address how Agrippa, the multimedia work of literature, is based on a mechanism of nostalgia, whose ‘remarkable staying power’ sprung from a photo-album, replicated into a codex-form book and found itself ultimately enclosed – according to technological obsolescence and library policies and procedures – in a 1990s relic: a diskette. ‘Hacked’ at Agrippa’s launch event, the diskette’s content was subsequently retyped and posted on the Internet, preserving access to Gibson’s poem – via digital surrogates – but also separating the text from Ashbaugh’s images, to date. To return to Barthes, if the continued online presence of the poem might convey the ‘temporal hallucination’ of an early electronic publication, I argue that the dual nature of the book – part-analogue, part-digital – has remained nostalgic to be found since 1992.

It is a known tension of queer digital ecosystems that visibility produces a difficult binary. Each advance in seamless communication and its multimodal tools allows queer individuals to express themselves in new, imaginative ways. In doing so, new sexualities and gendered ways of being emerge that expand and destabilize once rigidly held binaries. Yet each of these advances comes with an inevitable backlash in which content moderation conflates queerness with adult content, erasing it from platforms by decoupling disclosure from its contextual and communal formations. The result of this cyclical tension of in/visibility is a perception that queerness did not exist in the immediate past or in any formation of digital space. While queer internet scholars like Avery Dame-Griff and Kat Brewster have worked tirelessly to correct such perceptions, the sentiment produces a sociohistorical nostalgia for artifacts of analog circulation. Ironically, much of the access to these artifacts comes through digital repositories of queer history such as the Digital Transgender Archive and the GLBT Historical Society. However, as contemporary historical evidence, these records can be informationally biased. Tools like OCR can inadvertently obfuscate the technologies of circulation that helped put queerness online. Put differently, we know how queerness moved from newsletters to early bulletin board systems, but we know much less about how queer folks used information and communication technologies to enact that onboarding. It is through examining these interactions that this presentation will explore the relationships between informational nostalgia and obsolete media and, in doing so, contend that queerness has a generative relationship with obsolescence. Borrowing from queer theoretical lenses such as Halberstam’s notion of queer failure and Munoz’s theories of queer archival ephemera, this presentation re-examines a handful of proto-Internet queer media artifacts, ranging from transgender newsletters produced during the rise of the Internet to subscription-based VHS tapes released at the decline of magnetic media. In highlighting these artifacts, the presentation asks what they reveal about the informational desires of their contemporary users and how those desires might serve as a lifeline of nostalgia for queer people today, hoping to build queer information networks anew.

Toggle Content

Respondents: Tess McNulty and Matthew Kirschenbaum

Approximately 1:50 pm EST / 20:50 CET / 12:50 CST 11:50 AM PST

Conference End

No later than 2:30 PM EST / 21:30 CET / 1:30 CST / 12:30 PM PST