{"id":893,"date":"2022-05-15T14:56:40","date_gmt":"2022-05-15T13:56:40","guid":{"rendered":"http:\/\/databaz.org\/xtrm-art\/?p=893"},"modified":"2022-05-15T15:01:49","modified_gmt":"2022-05-15T14:01:49","slug":"recycling-dialectique-de-l%e2%80%99image-dechet","status":"publish","type":"post","link":"https:\/\/databaz.org\/xtrm-art\/?p=893","title":{"rendered":"Recycling : dialectique de l\u00e2\u20ac\u2122image-d\u00c3\u00a9chet"},"content":{"rendered":"<p>&nbsp;<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" src=\"http:\/\/databaz.org\/xtrm-art\/wp-content\/uploads\/image\/fond_texte.jpg\" alt=\"\" width=\"1000\" height=\"1006\" align=\"textTop\" \/><\/p>\n<p>Tous les jours 350 millions de photos sont publi&eacute;s sur facebook. Plus de 240 milliards d&rsquo;images ont &eacute;t&eacute; ainsi accumul&eacute;es par la plateforme depuis sa cr&eacute;ation. Les r&eacute;seaux sociaux ont acc&eacute;l&eacute;r&eacute; &agrave; la fois la production des images num&eacute;riques et d&rsquo;autre par leur publication, leur partage, leur diffusion, leur duplication.<br \/>\nDans leur article publi&eacute; dans <a href=\"https:\/\/aoc.media\/opinion\/2022\/04\/05\/le-blanchiment-des-images\/\">AOC<\/a>, Gwenola Wagon et St&eacute;phane Degouttin soulignent avec pertinence en quel sens les imaginaires des images diffus&eacute;es, de l&rsquo;open-space &agrave; l&rsquo;image &eacute;rotico-porno, du portrait &agrave; la vue d&rsquo;ensemble, correspondent &agrave; des imaginaires reli&eacute;s &agrave; des stocks d&rsquo;image, &agrave; &laquo;&nbsp;une hygi&egrave;ne de l&rsquo;image&nbsp;&raquo;, &laquo;&nbsp;une strat&eacute;gie d&rsquo;&eacute;vidage de l&rsquo;image&nbsp;&raquo;. Ainsi les images viendraient recouvrir le r&eacute;el par leur esth&eacute;tique clean, par leur codification apriori, effa&ccedil;ant les asp&eacute;rit&eacute;s du r&eacute;el, et ce qu&rsquo;elles engagent comme probl&eacute;matiques (sociales, politiques, etc) en faveur d&rsquo;une vision &eacute;dulcor&eacute;e, adoucie, anesth&eacute;si&eacute;e.<\/p>\n<p>Se pose cependant la question de savoir pour qui sont destin&eacute;es ces images ? En quel sens, les r&eacute;seaux sociaux demandent que soient post&eacute;es des images, sachant que les images stock&eacute;es exigent de la place sur des servers donc des co&ucirc;ts r&eacute;els et de la place mat&eacute;rielle.<\/p>\n<p>Mon hypoth&egrave;se, c&rsquo;est que l&rsquo;image prise dans une dialectique qui d&eacute;finit sa fin, ne correspond plus r&eacute;ellement &agrave; ce qu&rsquo;elle &eacute;tait traditionnellement, mais qu&rsquo;elle est imm&eacute;diatement autre chose, &agrave; savoir un d&eacute;chet qui est con&ccedil;u m&eacute;diatement pour &ecirc;tre recycl&eacute;.<\/p>\n<p>Intentionnellement, du point de vue de la conscience humaine, lorsque l&rsquo;on poste une image, c&rsquo;est pour  qu&rsquo;elle soit vue. C&rsquo;est ce qui tient au d&eacute;sir. Et c&rsquo;est ce d&eacute;sir de reconnaissance m&eacute;diatis&eacute;e de soi que stimulent les strat&eacute;gies relationnelles des r&eacute;seaux sociaux. J&rsquo;appelle la profondeur de l&rsquo;image, ce qui am&egrave;ne une image &agrave; nous retenir dans son temps, dans ses plis, dans sa mat&eacute;rialit&eacute;, dans son appara&icirc;tre (aiesthesis). La profondeur d&rsquo;une image n&rsquo;est pas ainsi une dimension mat&eacute;rielle, mais temporelle.<\/p>\n<p>Ce que le cr&eacute;ateur d&rsquo;une image, par exemple d&rsquo;un tableau peut souhaiter c&rsquo;est que son image retienne l&rsquo;attention. La dimension du regard tient au temps et c&rsquo;est dans ce temps que l&rsquo;image affecte.<\/p>\n<p>Or, les r&eacute;seaux sociaux ont d&eacute;velopp&eacute; une logique de flux tendu pour les informations ou images qui leur sont envoy&eacute;es.  L&rsquo;image n&rsquo;est plus pens&eacute;e selon une logique de profondeur mais de transmissibilit&eacute; et de stimuli rapides. Ce n&rsquo;est plus la qualit&eacute; de l&rsquo;image qui est primordiale, mais la quantit&eacute; de partage et de like. Le like et le partage ne reposent pas sur une analyse qualitative mais sur d&rsquo;autres logiques li&eacute;es &agrave; des processus algorithmiques : la fr&eacute;quence de post, le r&eacute;seau interconnect&eacute; avec celui qui poste.  D&egrave;s lors l&rsquo;image est pens&eacute;e en tant que potentialit&eacute; de like et elle exige imm&eacute;diatement d&rsquo;&ecirc;tre d&eacute;pass&eacute;e dans la prochaine image post&eacute;e. La reconnaissance n&rsquo;existe que dans l&rsquo;enchainement des publications. Les publications ne sont pas pens&eacute;es comme li&eacute;es, mais &agrave; chaque fois comme des objets autonomes. Cela se voit aussi bien par des traits sp&eacute;cifiques dans les cr&eacute;ations vid&eacute;os : des mini-sc&eacute;narios, d&rsquo;une dur&eacute;e courte, sans lien avec les pr&eacute;c&eacute;dentes cr&eacute;ations. Les r&eacute;seaux sociaux sont ainsi des flux de points qui doivent se succ&eacute;der ind&eacute;finiment. L&rsquo;&eacute;paisseur temporelle doit &ecirc;tre r&eacute;duite le plus possible.<\/p>\n<p>La logique de flux fait, et ici nous je rejoins totalement GW SD que l&rsquo;image doit correspondre avec une esth&eacute;tique du r&eacute;seau. Cette logique de mim&eacute;tisme, de m&egrave;me (Dawkins), est une des cl&eacute;s pour r&eacute;ussir &agrave; augmenter la fr&eacute;quence de like et de transmissibilit&eacute; d&rsquo;une image. Se d&eacute;veloppe ainsi une forme d&rsquo;acad&eacute;misme de l&rsquo;image, de modalit&eacute; de la r&eacute;p&eacute;tition dans une diff&eacute;rence qui doit se nier d&rsquo;une certaine mani&egrave;re.<\/p>\n<p>L&rsquo;image n&rsquo;attire plus l&rsquo;attention, elle entre en &eacute;cho avec un d&eacute;j&agrave; vu.<\/p>\n<p>Ce que j&rsquo;appelle l&rsquo;image-d&eacute;chet, est ainsi l&rsquo;image con&ccedil;ue selon l&rsquo;intentionnalit&eacute; des promoteurs des r&eacute;seau sociaux non plus comme devant avoir une forme d&rsquo;&eacute;paisseur temporelle, mais &eacute;tant imm&eacute;diatement pens&eacute;e selon sa logique de recyclage. Le but des r&eacute;seaux sociaux tient &agrave; l&rsquo;acc&eacute;l&eacute;ration des publications, en vue d&rsquo;un stockage-recyclage de donn&eacute;es.<\/p>\n<p>Nous le savons, nos informations, qu&rsquo;elles soient textuelles, picturales, vid&eacute;os ou audios, entrent toutes dans des bases de donn&eacute;es (big data) qui elles-m&ecirc;mes servent &agrave; alimenter des processus d&rsquo;apprentissage pour des intelligences artificielles.<\/p>\n<p>Le but des r&eacute;seaux sociaux n&rsquo;est pas l&rsquo;attention port&eacute;e par les humains. Cette attention n&lsquo;est qu&rsquo;une m&eacute;diation, en vue d&rsquo;une autre fin. L&rsquo;attention humaine doit m&ecirc;me &ecirc;tre la plus courte possible pour une image, quelqu&rsquo;elle soit. En effet nous pourrions imaginer d&rsquo;une mani&egrave;re absurde, une image qui retiendrait absolument l&rsquo;attention, qui impossibiliserait toute autre image, cela an&eacute;antirait toute logique de r&eacute;seau social. Nous serions alors dans une perspective &agrave; la Borg&egrave;s plus qu&rsquo;&agrave; la Meta ou la Google. <br \/>\nAu contraire les images doivent &ecirc;tre suffisamment volatiles, pour qu&rsquo;elles se succ&egrave;dent sans que nous nous y arr&ecirc;tions. Certes, strat&eacute;giquement, il est toujours utile pour cette logique de mettre en &eacute;vidence des flux qui se d&eacute;marquent. Mais ils ne sont publicitairement con&ccedil;us que comme des stimuli pour inciter tout un chacun &agrave; poster.<\/p>\n<p>Les images ne nous sont plus vraiment adress&eacute;es. Elles ne sont plus destin&eacute;es au regard humain, mais elles sont envoy&eacute;es &agrave; des IA. L&agrave; aussi l&rsquo;image n&rsquo;est pas pens&eacute;e dans une logique d&rsquo;hyper-production comme r&eacute;alit&eacute;, mais elle est m&eacute;diation pour saisir &agrave; partir d&rsquo;intelligences artificielles, des informations qui ne concernent pas l&rsquo;image en propre. L&rsquo;image-d&eacute;chet est ainsi une image qui sert de mati&egrave;re &agrave; recycler pour des strat&eacute;gies aussi bien commerciales, politiques, sociologiques qu&rsquo;esth&eacute;tique. Ce qui compte n&rsquo;est pas l&rsquo;image en tant qu&rsquo;entit&eacute;, mais l&rsquo;image en tant que multiplicit&eacute; de donn&eacute;es exploitables.<\/p>\n<p>Lorsque l&rsquo;on cr&eacute;e avec une IA (par exemple les syst&egrave;mes CLIP), la cr&eacute;ation de l&rsquo;image est le r&eacute;sultat du recyclage de nos images qui ont &eacute;t&eacute; inject&eacute;es dans des datasets et analys&eacute;es gr&acirc;ce &agrave; du deep learning. La cr&eacute;ation &agrave; l&rsquo;aide d&rsquo;IA se caract&eacute;rise ainsi comme une forme qui se constitue &agrave; partir de potentialit&eacute;s qui sont tir&eacute;es de l&rsquo;image-d&eacute;chet. Ce que nous voyons comme r&eacute;sultat est ainsi tout &agrave; la fois ressemblant (car issu de donn&eacute;es &eacute;l&eacute;mentaires d&rsquo;images pr&eacute;-existantes) et dissemblabes (au sens o&ugrave; l&rsquo;IA n&rsquo;est pas dans une mim&eacute;sis de perception, mais dans une mim&eacute;sis par statistique &agrave; partir  d&rsquo;analyses de donn&eacute;es).<\/p>\n<p>&nbsp;<\/p>\n<p>_______________________________________________<\/p>\n<p>&nbsp;<\/p>\n<p>Every day 350 million photos are published on facebook. More than 240 billion images have been accumulated by the platform since its creation. Social networks have accelerated both the production of digital images and others through their publication, sharing, dissemination, duplication.<\/p>\n<p>&nbsp;<\/p>\n<p>In their article published in AOC, Gwenola Wagon and St&eacute;phane Degouttin pertinently underline the sense in which the imaginaries of the images disseminated, from the open-space to the erotic-porn image, from the portrait to the overview, correspond to imaginaries related to image stocks, to &quot;an image hygiene&quot;, &quot;a strategy of hollowing out the image&quot;. Thus the images would come to cover the real by their clean aesthetics, by their a priori codification, erasing the roughness of the real, and what they involve as problematic (social, political, etc.) in favor of a watered down, softened, anesthetized vision.<\/p>\n<p>&nbsp;<\/p>\n<p>However, the question arises, for whom are these images intended? In what sense, social networks require images to be posted, knowing that the stored images require space on servers and therefore real costs and material space.<\/p>\n<p>&nbsp;<\/p>\n<p>My hypothesis is that the image taken in a dialectic that defines its end, no longer really corresponds to what it traditionally was, but that it is immediately something else, namely a waste that is mediately designed to be recycled.<\/p>\n<p>&nbsp;<\/p>\n<p>Intentionally, from the point of view of human consciousness, when we post an image, it is to be seen. This is what comes down to desire. And it is this desire for mediated self-recognition that the relational strategies of social networks stimulate. I call the depth of the image, which leads an image to retain us in its time, in its folds, in its materiality, in its appearing (aiesthesis). The depth of an image is thus not a material dimension, but a temporal one.<\/p>\n<p>&nbsp;<\/p>\n<p>What the creator of an image, for example of a painting, may wish is that his image attract attention. The dimension of the gaze depends on time and it is in this time that the image affects.<\/p>\n<p>&nbsp;<\/p>\n<p>However, social networks have developed a tight flow logic for the information or images sent to them. The image is no longer conceived according to a logic of depth but of transmissibility and rapid stimuli. It is no longer the quality of the image that is essential, but the quantity of shares and likes. The like and the sharing are not based on a qualitative analysis but on other logics linked to algorithmic processes: the frequency of posting, the network interconnected with the one who posts. From then on the image is thought of as a potentiality of like and it immediately demands to be exceeded in the next image posted. Recognition only exists in the chain of publications. The publications are not thought of as linked, but each time as autonomous objects. This can also be seen by specific features in the video creations: mini-scenarios, of short duration, unrelated to previous creations. Social networks are thus flows of points that must follow one another indefinitely. The temporal thickness must be reduced as much as possible.<\/p>\n<p>&nbsp;<\/p>\n<p>The flow logic does, and here we totally agree with GW SD that the image must correspond with the aesthetics of the network. This logic of mimicry, of meme (Dawkins), is one of the keys to successfully increasing the frequency of likes and the transmissibility of an image. Thus develops a form of academicism of the image, of modality of repetition in a difference that must be denied in a certain way.<\/p>\n<p>&nbsp;<\/p>\n<p>The image no longer attracts attention, it echoes with a deja vu.<\/p>\n<p>&nbsp;<\/p>\n<p>What I call the image-waste, is thus the image conceived according to the intentionality of the promoters of the social networks no longer as having to have a form of temporal thickness, but being immediately thought out according to its logic of recycling. The purpose of social networks is to accelerate publications, with a view to storing and recycling data.<\/p>\n<p>&nbsp;<\/p>\n<p>We know that our information, whether textual, pictorial, video or audio, all enters databases (big data) which are themselves used to feed learning processes for artificial intelligence.<br the=\"\" purpose=\"\" of=\"\" social=\"\" media=\"\" is=\"\" not=\"\" human=\"\" attention.=\"\" this=\"\" attention=\"\" only=\"\" a=\"\" with=\"\" view=\"\" to=\"\" another=\"\" end.=\"\" must=\"\" even=\"\" be=\"\" as=\"\" short=\"\" possible=\"\" for=\"\" an=\"\" whatever=\"\" it=\"\" is.=\"\" indeed=\"\" we=\"\" could=\"\" imagine=\"\" in=\"\" absurd=\"\" image=\"\" that=\"\" would=\"\" absolutely=\"\" hold=\"\" which=\"\" disable=\"\" any=\"\" other=\"\" destroy=\"\" logic=\"\" network.=\"\" then=\"\" borg\u00c3\u00a8s=\"\" perspective=\"\" more=\"\" than=\"\" meta=\"\" or=\"\" \/><br \/>\nOn the contrary, the images must be sufficiently volatile, so that they follow one another without our stopping there. Certainly, strategically, it is always useful for this logic to highlight flows that stand out. But they are only advertised as stimuli to encourage everyone to post.<\/p>\n<p>&nbsp;<\/p>\n<p>The images are no longer really addressed to us. They are no longer intended for the human gaze, but they are sent to AIs. Here too the image is not thought of in a logic of hyper-production as reality, but it is mediation to seize from artificial intelligence, information that does not concern the image itself. The waste-image is thus an image that serves as material to be recycled for commercial, political, sociological and aesthetic strategies. What counts is not the image as an entity, but the image as a multiplicity of exploitable data.<\/p>\n<p>&nbsp;<\/p>\n<p>When creating with an AI (for example with CLIP systems), the creation of the image is the result of the recycling of the broadcast images that have been injected into datasets and analyzed using deeplearning. Creation using AI is thus characterized as a form that is constituted from potentialities that are drawn from the waste-image. What we see as a result is thus both resembling (because stemming from elementary data of pre-existing images) and dissimilar (in the sense that the AI \u00e2\u20ac\u2039\u00e2\u20ac\u2039is not in a mimesis of perception, but in a mimesis by statistics from data analysis).<\/p>\n<p>&nbsp;<\/p>\n","protected":false},"excerpt":{"rendered":"<p>&nbsp; Tous les jours 350 millions de photos sont publi&eacute;s sur facebook. Plus de 240 milliards d&rsquo;images ont &eacute;t&eacute; ainsi accumul&eacute;es par la plateforme depuis sa cr&eacute;ation. Les r&eacute;seaux sociaux ont acc&eacute;l&eacute;r&eacute; &agrave; la fois la production des images num&eacute;riques et d&rsquo;autre par leur publication, leur partage, leur diffusion, leur duplication. Dans leur article publi&eacute; [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[11,135,7,6],"tags":[137,195,187,47,196,136,193,234,194],"class_list":["post-893","post","type-post","status-publish","format-standard","hentry","category-article","category-intelligence-artificielle","category-texte","category-une","tag-ai","tag-aoc","tag-dechets","tag-digital-art","tag-gwenola-wagon","tag-ia","tag-image","tag-intelligence-artificielle","tag-stephane-degoutin"],"_links":{"self":[{"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=\/wp\/v2\/posts\/893","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=893"}],"version-history":[{"count":3,"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=\/wp\/v2\/posts\/893\/revisions"}],"predecessor-version":[{"id":897,"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=\/wp\/v2\/posts\/893\/revisions\/897"}],"wp:attachment":[{"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=893"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=893"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/databaz.org\/xtrm-art\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=893"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}