Source-linked AI summary
Latent-surrealism: Revisiting surrealism and its aesthetics in relation to contemporary AI-Generated cultural production
Anca-Simona Horvath
TL;DR
The chapter asks how AI-generated creative objects relate to Dadaism, Surrealism, and Conceptual Art. Through Carroll’s anti-rationalist literary world and three AI-generated architecture case studies, it develops latent-surrealism as a framework for AI-media aesthetics and methods. It argues that natural-language prompting shifts creative production toward language-centered practice.
Problem
AI-media objects are reshaping cultural production, but their aesthetic and methodological relation to earlier avant-garde movements requires theoretical examination.
Method
The chapter relates Dadaism, Surrealism, and Conceptual Art to Carroll’s Through the Looking-Glass and analyzes Deep House, Post-Pharaonic Architecture, and A State of Decay.
Results
The chapter develops latent-surrealism to describe AI-media conditions including readymade-data reassembly, machine absurdity, and craft’s decoupling from artistic value.
Takeaways & Limitations
Natural language and prompting become central creative media, marking a linguistic turn in fields that use AI-media objects.
Abstract
from arXiv · showhide
This chapter examines AI-media objects - creative and artistic outputs generated through generative artificial intelligence in the form of text-to-X tools - in relation to three avant-garde movements of the twentieth century: Dadaism, Surrealism, and Conceptual Art. Drawing on Lewis Carroll's Through the Looking-Glass as an early precursor to these three movements and to anti-rationalist aesthetics, and on three case studies in AI-generated conceptual architecture - Matias del Campo's "Deep House" and Hassan Ragab's "Post-Pharaonic Architecture" and "A State of Decay" - the chapter develops the concept of latent-surrealism. Latent-surrealism includes a set of aesthetic and methodological conditions inherent to creative AI-media objects. These include the use of readymade datasets reassembled through collage-like processes, the absurd as an aesthetic quality of machine hallucinations, and the decoupling of craft from artistic value. The chapter further argues that AI-media objects represent a shift in the conditions of creative production: where earlier computational tools required graphical interfaces and programming literacy, natural language now functions as the operative medium. This repositions the prompt (as language-based instructions) in the center of the creative process, in continuity with the legacy of Conceptual Art, and signals a linguistic turn across creative fields that make (use of) AI-media objects.
Introduction
The chapter situates AI-media within creative practice while emphasizing its difference from earlier computational tools. It outlines case studies used to develop latent-surrealism and examine language-centered AI production.
- Generative AI differs from earlier creative computational tools by using natural language to produce text, images, video, and audio.Earlier tools relied on graphical interfaces or programming languages and often required familiarity with computational thinking.
- The analysis examines how AI-media uses language and how prompting becomes central to creative production.
- The chapter develops latent-surrealism through three AI-generated conceptual architecture case studies: Deep House, Post-Pharaonic Architecture, and A State of Decay.
Dada and Surrealism as Artistic Movements
Dada and Surrealism established anti-rationalist artistic programs that challenged conventional assumptions about art, while Conceptual Art elevated ideas and language over craft and physical form.
- Dada emerged in 1916 as an international anti-rational and anti-art movement founded by artists and writers in exile.
- The readymade shifted artistic value from skill, intentionality, and formal transformation toward designation and concept.
- Surrealism inherited Dada’s anti-rationalism but redirected it toward liberating the unconscious through a positive cultural programme.
- Conceptual Art extended these developments by privileging ideas over physical form and making language a primary medium of art-making.
The Chessboard: Surreal Worldbuilding in Through the Looking Glass
Carroll’s Through the Looking-Glass challenges rational order through dream logic, unstable language, and a mirror world where conventional causality breaks down.
- Carroll’s works challenged Victorian thought through language, dream logic, and the subversion of cause and effect.
- Through the Looking-Glass constructs a world where cause and effect do not follow one another and words acquire meanings through arbitrary choice.
- The chessboard evokes the Cartesian grid as a metaphor for Enlightenment order within the novel’s irrational mirror world.
Case Study 1: Matias del Campo, Deep House
Deep House uses AI-generated architectural form to produce estranged, uncanny spaces. Its design process treats the architect as an interlocutor steering outputs that remain surprising and partly beyond human control.
- Deep House applies estrangement to conceptual design for a single-family home whose forms are generated through unpredictable AI outputs.
- StyleGAN2 trained on over 2,000 annotated house plans generated forms recognizable as architecture yet unfamiliar as typologies.
- The architect works through prompts as an interlocutor, negotiating a generative process whose outcomes remain surprising and not fully controlled.
- Deep House maintains tension between the familiar and alien, and between rational and irrational internal logics, producing an uncanny architectural object.
Case Study 2: Hassan Ragab, Post-pharaonic Architecture and A State of Decay
Hassan Ragab uses iterative prompt-crafting to navigate AI latent space, generating architectural images through linguistic negotiation. His projects also expose how Western-biased training archives distort and assimilate non-Western architectural traditions.
- Ragab’s prompt-crafting iteratively navigates an AI model’s latent space toward desired aesthetic solutions.The process is described as a linguistic negotiation producing imagery neither practitioner nor model could independently produce.
- Post-Pharaonic Architecture combines ancient Egyptian architectural vocabulary with twentieth-century abstract expressionism.The resulting structures fragment and restack stone masses into compositions unrelated to Pharaonic constructional convention or expressionist spatial logic.
- A State of Decay applies latent interpolation to Art Nouveau facades, producing stone and glass surfaces that dematerialize architectural forms.
- Ragab’s work reveals how Western-dominated training archives distort or assimilate invoked Pharaonic and Islamic traditions into Western typologies.He frames this displacement as a form of latent archaeology conducted through prompt crafting.
Latent-Surrealism in AI-media
Latent-surrealism names the surrealist aesthetics and methods that the chapter argues are inherent in creative AI-media outputs. It connects latent-space generation with a shift from programming-centered computation toward natural language as a creative medium.
- The chapter defines latent-surrealism as inherent in creative outputs generated using AI-media.
- AI latent space encodes compressed, high-dimensional representations organized by learned similarity rather than Cartesian coordinates.Generative models sample from and interpolate within this space when producing new images.
- Earlier computational creative tools were described through programming, programmability, modularity, automation, variability, and transcoding.
- Text-to-X AI technologies make natural language an operative means of producing media across multiple creative fields.In architectural design, the chapter characterizes this development as a linguistic turn in a field dominated by visual representation.
- AI-media production involves prompts, programming languages, and annotations, each operating at a different layer of interaction and computation.
A Collage of Ready-mades
The chapter interprets AI-media as reassembling readymade dataset content through collage-like processes, producing hybrid forms that require authorial curation.
- Like Dadaist and Surrealist readymades and collages, AI-media reassembles large datasets of existing content into new hybrid forms.These hybrids can be understood as interpolations between distinct points in latent space and require complex authorial curation.
The Absurd
The chapter links absurd AI-generated hybrids to a shared aesthetic category of Dadaism, Surrealism, and Lewis Carroll’s work. It further presents machine hallucinations as possible creative materials.
- AI hybrids produced through latent-space blending are sometimes absurd, connecting machine generation to Dadaist and Surrealist aesthetics.The chapter also relates this absurdity to Carroll’s work.
- AI hallucinations are presented as creative materials within a broader lineage that includes glitch-based theories of machine creativity.
The Concept
Generative AI makes AI-media broadly accessible through simple natural-language instructions, shifting value toward concepts rather than execution or craftsmanship.
- Natural-language instructions let users create AI-media without mastering complex interfaces, computational thinking, or programming languages.The passage presents this accessibility as potentially democratizing production across video, image, and sound.
- Because media can be produced easily and cheaply, original ideas become more important than craft or virtuosity.Craftsmanship is described as encoded in the tools rather than supplied entirely by the maker.
The Techno-feudal
Generative AI is accessible but differs from earlier creative computational tools through paid, platform-dependent services and resource-intensive infrastructure.
- Generative AI is often monetized through paid subscriptions or limited free access that can make users dependent on proprietary platforms.This contrasts with earlier creative tools that were often free or available through one-off payments.
- The infrastructure behind accessible generative AI consumes vast amounts of energy and water, costs that are not immediately visible to users.The chapter connects this model to techno-feudalism and privatized internet infrastructure.
Conclusion
The chapter frames text-to-X creative production through Dadaism, Surrealism, and Conceptual Art, developing latent-surrealism as a way to analyze AI-media. It links accessible natural-language generation to collage, machine hallucination, and a reduced role for craft, while noting infrastructural and ecological costs.
- Conclusion: The chapter studies early-stage architectural design to distinguish text-to-X generative models from earlier software, programming languages, and computational thinking.Its case studies include Deep House, Post-pharaonic Architecture, and A State of Decay.
- Conclusion: Latent-surrealism describes AI-media through ready-made datasets, collage-like recombination, absurd hallucinations, and the reduced importance of craft.The chapter connects these conditions to Surrealism, Dada, and Conceptual Art.
- Conclusion: The chapter argues that AI-assisted artistic production places language at the center of creative work and makes the concept more important than execution.It also identifies paid services and energy- and water-intensive infrastructure as ethical and ecological concerns.