{"id":2065615,"date":"2026-07-03T11:47:42","date_gmt":"2026-07-03T09:47:42","guid":{"rendered":"https:\/\/www.luxsure.fr\/2026\/07\/03\/38-paintings-20000-images-what-generative-ai-is-challenging-in-the-luxury-industry\/"},"modified":"2026-07-03T11:50:13","modified_gmt":"2026-07-03T09:50:13","slug":"38-paintings-20000-images-what-generative-ai-is-challenging-in-the-luxury-industry","status":"publish","type":"post","link":"https:\/\/www.luxsure.fr\/en\/2026\/07\/03\/38-paintings-20000-images-what-generative-ai-is-challenging-in-the-luxury-industry\/","title":{"rendered":"38 paintings, 20,000 images: What Generative AI Is Challenging in the Luxury Industry"},"content":{"rendered":"\n

Vermeer painted 38 paintings. Media theorist Lev Manovich, on the other hand, generated and stored 20,000 images in two years, before narrowing them down to just a few hundred. Between these two figures lies a question that every fashion house must answer: where does scarcity begin when production costs nothing? <\/p>\n\n

The curator we didn’t hire<\/h2>\n\n

Manovich used Midjourney almost every day for two years. He produced hundreds of thousands of images, saved 20,000 of them, then narrowed his selection down to 4,000 reworked pieces, ultimately publishing only a few hundred on his website. Even this reduced number far exceeds the output of a twentieth-century artist, and even more so that of a master of the past: Bosch painted fewer than 50 canvases, and Vermeer 38. The painters who exhibited in Paris between 1870 and 1880\u2014among the Impressionists\u2014produced an average of nearly 1,000 works over the course of their careers, though many of them produced only between 400 and 800. The difference is not one of degree, but of nature: the artist equipped with a generative tool no longer creates; he or she curates. And curation, Manovich observes, becomes more difficult as abundance grows, because algorithmic production tends toward an almost uniformly constant aesthetic quality\u2014nothing truly stands out from the crowd. The phenomenon is not new: the programmed generative art of the 1960s and 1970s was already producing series of comparable quality based on random rules. What has changed is the scale\u2014and thus the cost of discernment. <\/p>\n\n

For a fashion house\u2019s leadership, the reality extends beyond the design studio. Any function that relies on the creation of images\u2014campaigns, content, variations on a piece\u2014faces the same paradox: scarcity, the historical driver of desirability, is no longer a constraint on production. It has once again become a management decision. <\/p>\n\n

The signature as an asset, not just for show<\/h2>\n\n

Generative tools separate what no other medium has ever separated so clearly: content and style. The same composition can be rendered as a watercolor, a 1980s film photograph, or in the style of a specific artist, without the meaning of the image changing. Manovich points out that the history of Western art has long been based on the opposite: a limited religious subject matter\u2014the Crucifixion, biblical scenes\u2014rendered in highly diverse styles, from Bosch to Bruegel, from Raphael to Watteau. It is style, rather than subject matter, that has earned these artists a place in the collections of the Louvre or the Centre Pompidou. If generative grammar makes this style interchangeable\u2014adjustable like a control parameter\u2014the question facing a fashion house is no longer merely aesthetic. It is one of heritage: who determines the brand\u2019s identity when any content can now be rendered in its visual form? <\/p>\n\n

The Details<\/strong> \u2014 Apple has added a feature to iOS 27 that converts photos into volumetric representations (Gaussian splats), allowing users to rotate the scene after the photo is taken. The company\u2019s teams have capped this rotation at 20% of the initial shooting angle\u2014an arbitrary threshold designed to distinguish between enhancing a memory and creating a new one. The decision is not technical; it is ethical: it establishes, in a single number, the boundary between memory and invention. To date, no other company has formalized the equivalent for its own generative tools. <\/p>\n\n

\"\"<\/figure>\n\n

A tool built for what’s expected<\/h2>\n\n

One final point, and the most counterintuitive one: generative AI, trained to predict the most likely continuation of a sequence, structurally produces the most expected image, not the most unique one. This is the opposite of the approach taken by the Parisian avant-garde 120 years ago\u2014Picasso, Braque, Soutine, Giacometti\u2014who sought precisely to defamiliarize the ordinary, or of the Surrealists, who championed the improbable image. The academic painters of the Salon, for their part, were already perfecting a polished, anatomically flawless aesthetic with sharp contours in the nineteenth century\u2014and were the most celebrated of their time before falling into obscurity. Generative AI, by its statistical nature, extends this conservative tradition rather than breaking with it. For an industry whose value rests on the exceptional, entrusting image production to a tool optimized for the probable is not a technical detail: it is a choice that defines the very nature of what is desirable. <\/p>\n\n

Manovich sums up his own experience with a question he considers poorly phrased, but one he continues to ask himself: what can be done that is radically new, given that a century of modern art and thirty years of computer-generated imagery have, in his view, already exhausted the bulk of what is visually possible? The real room for maneuver, he suggests, no longer lies in the image itself but in the interface that leads to it. It remains to be seen whether Maison\u2019s leadership will one day decide, for its own generative tools, to adopt the threshold Apple has set for the rotation of a memory\u2014or whether it will let that threshold set itself. <\/p>\n\n

\"\"<\/figure>\n","protected":false},"excerpt":{"rendered":"

Vermeer painted 38 paintings. Media theorist Lev Manovich, on the other hand, generated and stored 20,000 images in two years, before narrowing them down to just a few hundred. Between…<\/p>\n","protected":false},"author":2,"featured_media":2063287,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_lmt_disableupdate":"","_lmt_disable":"","footnotes":""},"categories":[77972],"tags":[78898,78895,78897,78896],"class_list":["post-2065615","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-luxury-and-ai","tag-ai-image-governance","tag-artificial-intelligence-and-art","tag-brand-desirability","tag-rarity-and-luxury"],"_links":{"self":[{"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/posts\/2065615","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/comments?post=2065615"}],"version-history":[{"count":0,"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/posts\/2065615\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/media\/2063287"}],"wp:attachment":[{"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/media?parent=2065615"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/categories?post=2065615"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.luxsure.fr\/en\/wp-json\/wp\/v2\/tags?post=2065615"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}