{"id":273517,"date":"2025-11-20T12:24:15","date_gmt":"2025-11-20T12:24:15","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/273517\/"},"modified":"2025-11-20T12:24:15","modified_gmt":"2025-11-20T12:24:15","slug":"retailers-liability-and-ai-applications-trademark","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/273517\/","title":{"rendered":"Retailers Liability And AI Applications &#8211; Trademark"},"content":{"rendered":"<p>The question of whether the user (consumer) or the&#13;<br \/>\nprovider of an AI application is liable for the infringement of&#13;<br \/>\nintellectual property rights as a result of the use of an&#13;<br \/>\nartificial intelligence application is an interesting one and&#13;<br \/>\nincreasingly relevant to retailers as they roll out more and more&#13;<br \/>\nAI applications. The recent Getty Images vs Stability AI decision&#13;<br \/>\nof the 4th November 2025 has shed a bit more light on&#13;<br \/>\nthis thorny issue.<\/p>\n<p>Liability of intermediaries: an old story<\/p>\n<p>The question of the level of liability of intermediaries in the&#13;<br \/>\nretail supply and promotion chain is not a new one. As referenced&#13;<br \/>\nin the Getty Images vs Stability AI decision, the Court of Justice&#13;<br \/>\nof the European Union (&#8216;CJEU&#8217;) has handed down judgments&#13;<br \/>\nwith regards the liability of search engine platforms such as&#13;<br \/>\nGoogle with regards keyword advertising in the well-known Google&#13;<br \/>\nFrance SARL and Google Inc. v Louis Vuitton Malletier SA&#13;<br \/>\n(C-236\/08), with regards the storage and dispatch of goods in Coty&#13;<br \/>\nGermany GmbH v Amazon Services Europe S\u00e0rl and others&#13;<br \/>\n(C\u2011567\/18) and with the regards the use of trade marks by&#13;<br \/>\nthird party advertisers in Daimler AG v Egy\u00fcd Garage&#13;<br \/>\nG\u00e9pj\u00e1rm\u0171jav\u00edt\u00f3 \u00e9s&#13;<br \/>\n\u00c9rt\u00e9kes\u00edt\u0151 Kft. (Case C\u2011179\/15). In&#13;<br \/>\nall these cases, the intermediary was held not liable for trade&#13;<br \/>\nmark infringement based essentially on defences of lack of active&#13;<br \/>\nknowledge of any infringing activity and the platforms\/parties&#13;<br \/>\nbeing essentially non-active in the infringing activity. In essence&#13;<br \/>\nthe parties were held not to be using the trade marks concerned in&#13;<br \/>\nthe course of trade, an essential pre-requisite for the finding of&#13;<br \/>\ntrade mark infringement.<\/p>\n<p>Now we come to the AI<\/p>\n<p>Now, we come to the question of the liability of AI applications&#13;<br \/>\nwhich is a new phenomenon to consider. Stability AI attempted to&#13;<br \/>\nuse the reasoning put forward in the Google France, Coty and&#13;<br \/>\nDaimler cases in the Getty case and argued that it was the user,&#13;<br \/>\nand the user alone, of its AI platforms who controlled the&#13;<br \/>\npotential output of infringing images created by them via the&#13;<br \/>\nprompts the user (consumer) entered into the AI applications.&#13;<br \/>\nStability AI was essentially arguing it was not liable; it was&#13;<br \/>\nmerely a tool not an active party in any infringing activity. Here&#13;<br \/>\nthe court was considering the question of trade mark infringement&#13;<br \/>\nbased on the production of the GETTY IMAGES and ISTOCK trade marks&#13;<br \/>\nin images produced by Stability AI applications. On this point,&#13;<br \/>\nStability AI lost.<\/p>\n<p>Counsel for Stability AI in cross examination with a witness&#13;<br \/>\nargued:<\/p>\n<p>&#8220;the model is a tool controlled by the user and the more&#13;<br \/>\ndetailed the prompt is, the more control is being&#13;<br \/>\nimposed.&#8221;<\/p>\n<p>In response the witness responded:<\/p>\n<p>&#8220;That is partially true. The user has control over what is&#13;<br \/>\nprompted. However, what the user does not have control over is what&#13;<br \/>\nthe model is trained on. We have no control over that. What the&#13;<br \/>\nuser does not have control over are any semantic guardrails that&#13;<br \/>\nmight be put on the prompt and any semantic guardrails that we put&#13;<br \/>\non the output. Absolutely, the user has control over what it asked&#13;<br \/>\nfor but does not have 100% control over what is coming out the&#13;<br \/>\nother end.&#8221;<\/p>\n<p>Stability AI lost this point because (a) it controlled the data&#13;<br \/>\non which its AI applications were trained and (b) in the Getty case&#13;<br \/>\nat least the judge seemed to be swayed by the fact that most users&#13;<br \/>\ndid not want GETTY IMAGES or ISTOCK trade mark appearing on their&#13;<br \/>\noutputted images, so it could be argued were actively not trying to&#13;<br \/>\nproduce infringing images.<\/p>\n<p>Now we come to the retailers and user&#8217;s&#13;<br \/>\nliability<\/p>\n<p>The exchange between Stability AI and the witness quoted above&#13;<br \/>\nfor me highlights some of the key issues of liability of AI&#13;<br \/>\napplications. A key issue seems to be control. It seems liability&#13;<br \/>\ncannot be pushed to the user, if the user has no absolute control&#13;<br \/>\nover the output of the AI application. Detailed prompts encouraging&#13;<br \/>\ninfringing activity, such as searching for counterfeit products,&#13;<br \/>\nmight make it more likely that liability could be pushed to user.&#13;<br \/>\nHowever, given that a user would rarely have complete control over&#13;<br \/>\nthe outputs of an AI application, after all that is the point of an&#13;<br \/>\nAI application.<\/p>\n<p>It seems that retailers might wish to concentrate more on&#13;<br \/>\nmitigating activities to counter infringing activities via their AI&#13;<br \/>\napplication than trying to push the buck to the user. So here we&#13;<br \/>\nhave to consider guardrails. It is interesting to note that&#13;<br \/>\nStability AI had put in place so-called &#8216;filtering&#13;<br \/>\nfunctionality&#8217; so the applications did not render&#13;<br \/>\nphoto-realistic likenesses of well-known public figures, which were&#13;<br \/>\nadded to combat concerns over fake news and propaganda. The prompts&#13;<br \/>\nwere scanned for celebrity names. So, it was at least aware of the&#13;<br \/>\npotential of some concerning activity via its AI models.<\/p>\n<p>So, what are the practical take aways for a retailer with an AI&#13;<br \/>\napplication in place? For a retailer to avoid liability as best it&#13;<br \/>\ncan and fall within the reasoning laid down in the Google France,&#13;<br \/>\nCoty and Daimler cases, it must in my eyes:<\/p>\n<p>&#13;<br \/>\nScreen the training data of the AI application to avoid&#13;<br \/>\ninfringing material.&#13;<br \/>\n&#13;<br \/>\nPut in place guardrails into the AI application to push users&#13;<br \/>\naway from infringing activity.&#13;<br \/>\n&#13;<br \/>\nPut in place takedown procedures to remove infringing activity&#13;<br \/>\nwhen it is brought to its attention.&#13;<\/p>\n<p>The above would push the retailer as much as possible into the&#13;<br \/>\ngroup of inactive enabler of commerce, akin to a keyword&#13;<br \/>\nadvertising platform such as Google, not an active participant in&#13;<br \/>\nany infringing activity. Afterall the Google, Coty and Daimler&#13;<br \/>\ncases all concerned some arguably infringing activity, but those&#13;<br \/>\nparties succeeded in their defences because they were considered&#13;<br \/>\nnot active in such activity.<\/p>\n<p>What the Getty v Stability AI case however does however seem to&#13;<br \/>\nmake clear is that once the AI application is deemed to using the&#13;<br \/>\ninfringing sign in the course of trade liability will rest with the&#13;<br \/>\nowner\/controller of the AI application, and that could in future be&#13;<br \/>\na retailer.<\/p>\n<p>The content of this article is intended to provide a general&#13;<br \/>\nguide to the subject matter. Specialist advice should be sought&#13;<br \/>\nabout your specific circumstances.<\/p>\n","protected":false},"excerpt":{"rendered":"The question of whether the user (consumer) or the&#13; provider of an AI application is liable for the&hellip;\n","protected":false},"author":2,"featured_media":24266,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[59,57,58,50,56,54,55],"class_list":["post-273517","post","type-post","status-publish","format-standard","has-post-thumbnail","category-united-kingdom","tag-gb","tag-great-britain","tag-greatbritain","tag-news","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/273517","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=273517"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/273517\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/24266"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=273517"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=273517"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=273517"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}