[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84407-en":3,"doc-seo-84407-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84407,7971461741311,"Ophelia","https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826",8,"Research & Report","Model Collapse: On Recursion, Noise, and Uncharted Machine Visions","Since 2023, warnings about model collapse highlight how AI-generated outputs contaminate training sets, triggering progressive degradation in model performance. While engineering frames collapse as a feedback-driven failure yielding repetition or pixel noise and eventual loss of coherence, the paper treats it as a recursive mirror tied to early analog video feedback experiments. Media-archaeological case studies link historical video synthesis and contemporary generative art to explain how recursive training exposes AI-generated data’s dependence on human-produced content, challenging transhumanist ideals.","Preprint for Ethics and Aesthetics of Artificial Intelligence (Naples: Orthotes Edizioni), October 2025.  \nModel Collapse:  \nOn Recursion, Noise, and Uncharted Machine Visions  \nViolaine Boutet de Monvel  \n[violaine.boutetdemonvel@biblhertz.it](violaine.boutetdemonvel@biblhertz.it 1)[ 1](violaine.boutetdemonvel@biblhertz.it 1)  \nBibliotheca Hertziana – Max Planck Institute for Art History MPRG Impett – Machine Visual Culture: Artificial Intelligence and the History of Seeing  \nAbstract:  \nSince 2023, computer scientists have warned against model collapse – the contamination of training sets with AI-generated outputs that progressively degrade model performance. Exemplifying a positive-feedback-driven failure, it produces effects such as word repetition or pixel noise, ultimately leading to a loss of meaning and coherence–at least from an engineering standpoint. From a creative one, however, collapse is not merely a breakdown: it also functions as a recursive mirror that recalls early analog video feedback experiments, raising once again the question of what happens when a system turns inward and sees itself. In such cases, socalled machine vision no longer transmits the world (as in tele-vision) but increasingly generates worlds from within. Drawing on media archaeology through case studies of both historical video synthesis techniques and contemporary artistic uses of machine learning, this paper examines what recursive training reveals about the dependent nature of AI-generated data. It argues that the potential effects of collapse challenge transhumanist ideals while inviting anaesthetic perspective, positioning noise and recursion as key concepts for understanding both artmaking and the AI ecosystem. Distributing agency across scales and networks, the latter currently remains reliant on new human-produced content, particularly within foundation models trained on massive datasets.  \nKeywords: recursive aesthetics, feedback, noise, video synthesis, generative AI.  \nIntroduction: Art as Noise Magnification  \nThe question of what model collapse reveals about the dependent nature of AI-generated data and its broader ecosystem must be situated within longer trajectories in the history of art, media, and cybernetics. My own point of entry arose from an attempt to define art in relation to engineering during my doctoral research at Sorbonne Nouvelle, titled From Video Feedback to Generative AI: On Recursivity in the Arts and Media (Boutet de Monvel, 2024) . That project built an aesthetic bridge between early video art, from the 1960s to the 1980s, and contemporary practices associated with generative AI, which have gained momentum since the mid-2010s due to breakthroughs in deep learning, beginning with GANs (Goodfellow, et al., 2014) and soon after diffusion models (Sohl-Dickstein, et al., 2015) . My aim was to reframe real-time analog synthesis – particularly video feedback techniques – as a precursor to current machinelearning-based art. To this end, I compared the room left for human mastery in the recursive operations of both closed-circuit video setups and generative AI models, which led me to articulate a networked, systemic view of artmaking that foregrounds technological agency alongside human intent, among other forces.  \nThis approach culminated in a definition, which I first proposed in 2021 during a seminar at Sorbonne Nouvelle, and later published in NECSUS: art strives to magnify the noise within and  \n[1](1 Correspondence: violaine.bdm@gmail.com)[ Correspondence: ](1 Correspondence: violaine.bdm@gmail.com)[violaine.bdm@gmail.com](1 Correspondence: violaine.bdm@gmail.com).  \nPreprint for Ethics and Aesthetics of Artificial Intelligence (Naples: Orthotes Edizioni), October 2025.  \nvia any given medium, as opposed to engineering that seeks to suppress it (Boutet de Monvel, 2023, p. 121) . In this context, noise refers to any agency that may introduce interference into the signal path beyond the artist’s design – for ","cbCaiiPQGoRJq0rS","https://ap.wps.com/l/cbCaiiPQGoRJq0rS","pdf",10058385,1,12,"English","en",105,"# Abstract\n# Introduction: Art as Noise Magnification\n## Defining art through recursion and engineering\n## Noise as interference beyond artistic design","[{\"question\":\"What does model collapse mean in the paper’s context?\",\"answer\":\"Model collapse is described as contamination of training sets with AI-generated outputs that create a positive-feedback loop, progressively degrading performance and producing effects like repetition or pixel noise.\"},{\"question\":\"How does the paper reinterpret collapse beyond an engineering failure?\",\"answer\":\"Instead of only treating collapse as malfunction, it frames collapse as a recursive mirror that echoes early analog video feedback experiments and raises questions about systems turning inward to generate worlds.\"},{\"question\":\"What is the paper’s key concept of “noise” and how does it relate to art?\",\"answer\":\"Noise is defined as agency that introduces interference into the signal path beyond the artist’s design, such as contingency from audiences or environments. The paper argues that art can magnify this noise by steering and orchestrating it rather than purely suppressing it, using this to analyze recursion like model collapse.\"}]",1784195390,30,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"model-collapse-on-recursion-noise-and-uncharted-machine-visions","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/model-collapse-on-recursion-noise-and-uncharted-machine-visions/84407/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What does model collapse mean in the paper’s context?","Question",{"text":75,"@type":76},"Model collapse is described as contamination of training sets with AI-generated outputs that create a positive-feedback loop, progressively degrading performance and producing effects like repetition or pixel noise.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the paper reinterpret collapse beyond an engineering failure?",{"text":80,"@type":76},"Instead of only treating collapse as malfunction, it frames collapse as a recursive mirror that echoes early analog video feedback experiments and raises questions about systems turning inward to generate worlds.",{"name":82,"@type":73,"acceptedAnswer":83},"What is the paper’s key concept of “noise” and how does it relate to art?",{"text":84,"@type":76},"Noise is defined as agency that introduces interference into the signal path beyond the artist’s design, such as contingency from audiences or environments. The paper argues that art can magnify this noise by steering and orchestrating it rather than purely suppressing it, using this to analyze recursion like model collapse.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":28,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]