Brussels Sprouts into the Global Market: The EU AI Act’s Reach into AI-Generated Advertisements

Of increasing significance in today’s digital and consumer-driven world is the use of artificial intelligence (AI) in advertisements globally. The European Union Artificial Intelligence Act (EU AI Act) (hereinafter, “the Act”) is the EU’s central law governing the use of AI and regulates parties based on how they use AI systems.[i] While companies within the EU are already subject to the Act, this post discusses how parties that use AI systems even outside of the EU, but whose “outputs” ultimately end up in the EU, are also subject to the Act.[ii] Thus, for example, global companies that use AI to create their advertisements (the “outputs”), where those advertisements are then used in the EU, must also comply with the Act.[iii] Failure to do so may result in hefty fines: The AI systems that companies use for creating advertisements would likely be categorized as “Limited Risk,” under which fines reach €15,000,000 or 3% of worldwide annual turnover for non-compliance.[iv]

Article 50 of the Act, “Transparency Obligations for Providers and Deployers of Certain AI Systems,” requires that companies[v] inform users if they are interacting with AI systems, unless it is obvious to a “reasonably well-informed, observant and circumspect” person considering the context and circumstances.[vi] Importantly, a company must indicate in a machine-readable and detectable format if it used AI to generate synthetic audio, image, video, or text content,[vii] which must be communicated to users in a clear and distinguishable manner.[viii] Thus, where a company’s AI-created advertisements—whether they be American or otherwise—end up in the EU, that company must disclose conspicuously to consumers that it used AI.[ix]

Recently, an advertisement by the clothing brand Guess stirred up controversy for using AI in Vogue's August 2025 print magazine.[x] The advertisement contained AI-generated fashion models, which, to the naked eye, may not have appeared obvious.[xi] However, the advertisement was labeled as being AI-generated, but this disclaimer was small and subtle[xii]—which arguably is not clear and distinguishable, as the Act would require.[xiii] Nevertheless, the campaign was featured in Vogue U.S.’s print edition,[xiv] not a European or digital version. Though, what if this print edition made its way physically to the EU—which is not inconceivable given Vogue’s strong European presence—or, what if, more likely, EU Vogue posted the advertisement to their social media pages?

Just weeks after the Vogue advertisement, clothing company J.Crew supposedly, though not confirmedly, used generative AI in a campaign posted to their Instagram account.[xv] A style investigator published an article noting numerous suspicious quirks in the advertisements, suggesting that they were created with AI—though, J.Crew did not disclose any AI use.[xvi] The investigator found the creator of the images, who is an AI photographer,[xvii] because his agency credited him with the images on their website[xviii] and LinkedIn (with the caption “#AI”).[xix] In response to the criticism, J.Crew reposted the advertisements, crediting the AI photographer with the “Digital art,” but without stating that it was AI-created (emphasis added).[xx] Though, importantly, the EU AI Act would require that J.Crew conspicuously made that known.[xxi]

Seemingly, where a brand posts an AI-generated advertisement on their social media account, as opposed to, for example, a physical advertisement directed only at an audience in their home country, it is possible that those advertisements could always be subject to the Act, given that social media cuts across international borders. [xxii] As such, in between now and August 2026, when Article 50 goes into effect,[xxiii] companies around the world will have to navigate the Act’s stringent requirements. However, it is likely that many will not be aware or informed that they must comply if they are outside of the EU. Compared to its western nation counterparts, the United States, at least on a federal level, lacks a strong, comprehensive regulation such as the EU AI Act.[xxiv] However, because AI has and will continue to become so advanced and commonly used globally, it is imperative that the U.S. and nations globally enact legislation that mirrors the Act in order to protect their citizens from unintentionally violating the EU AI Act.

Olivia A. Lilley is a staff member of Fordham International Law Journal Volume XLVIII. 

[i] Article 3 of the Act defines an AI system as: a “machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.” See id. art. 3.

[ii] See id. art. 2(1)(c). For further guidance on the scope of the Act, see Steven Farmer, et. al., The EU’s AI Act: A Review of the World’s First Comprehensive Law on Artificial Intelligence and What This Means for EU and Non-EU Companies, pillsbury law (July 19, 2024), https://www.pillsburylaw.com/en/news-and-insights/eu-ai act.html#:~:text=Providers%20must%20ensure%20that%20AI,manipulated%20(subject%20to%20exceptions.

[iii] See supra note 1, art. 2(1)(c).

[iv] The Act employs a risk-based framework, whereby the obligations imposed on parties depend on the risk levels of the AI systems being used. See supra note 1; see also id. art. 99 for penalties.

[v] See id. art. 50(1). Specifically, the Act applies to “Providers” (which includes companies), which is defined as: “a natural or legal person, public authority, agency or other body that develops an AI system or a general-purpose AI model or that has an AI system or a general-purpose AI model developed and places it on the market or puts the AI system into service under its own name or trademark, whether for payment or free of charge.” Id. art. 3(3). The Act also applies to “Deployers,” which is a party who uses an AI system under its authority except where the AI system is used in the course of a non-professional, personal activity. See id. art. (3)(4).

[vi] See id. art. 50(1). However, the Act does not apply to certain AI systems used for law enforcement. See id. art. 50(1).

[vii] See Id. art. 50(2).

[viii] See Id.

[ix] See Id.

[x] See Yasmin Rufo, Does this look like a real woman? AI model in Vogue raises concerns about beauty standards, bbc (July 26, 2025), https://www.bbc.com/news/articles/cgeqe084nn4o. While companies globally are utilizing the breadth of tools and services that AI has to offer for efficiency and cost-saving purposes, interestingly, the creators of the Guess advertisement charge in the low six-figures for brands like Guess.

[xi] See id.

[xii] See id.

[xiii] See supra note 9.

[xiv] See supra note 10.

[xv] See Blackbird Spylane, J.Crew used A.I. to counterfeit their own vibes, Substack: Blackbird Spylane (Aug. 26, 2025), https://www.blackbirdspyplane.com/p/jcrew-used-ai-to-counterfeit-their-own-vibes; see also Emily Leibert, et. al., J.Crew’s New Campaign Looks a Lot Like AI Slop, The Cut (Aug. 29, 2025), https://www.thecut.com/article/did-jcrew-use-ai-on-instagram-campaign.html.

[xvi] See id.

[xvii] See id.

[xviii] See CARTEL and Co., Sam Finn, J.Crew x Vans, CARTEL and Co., https://www.cartelandco.com/work/samfinnjcrewxvans.

[xix] See CARTEL and Co., LinkedIn, https://www.linkedin.com/posts/cartel-and-co_jcrew-vans-samfinn-activity-7360719391954563072-YBcn?utm_source=share&utm_medium=member_desktop&rcm=ACoAACxDMmMBIiTePU8oP96eWhZ8sfBxzm3iYQI (last visited Oct. 5, 2025).

[xx] See J.Crew (@jcrew), Instagram (August 9, 2025), https://www.instagram.com/p/DNIqS7-uBbR/?utm_source=ig_embed&ig_rid=9be04a94-9ec3-4feb-b143-33927a6bc3c7. Not only did J.Crew not directly acknowledge using AI, but even responded to The Cut’s request for comment with, “We’re always exploring new forms of creative expression, expanding how we work, and finding fresh, innovative ways to create content. This partnership with Sam Finn Studio is one of many examples of how we engage artists of all genres to interpret our brand and experiment with different art mediums.” See Leibert, supra note 15. An arguably well-executed crisis management response, but one that still refuses to admit the use of AI.

[xxi] See supra note 9, art. 50(9).    

[xxii] Though, what if an AI-generated advertisement is so well done that it is nearly indistinguishable from a non-AI advertisement? In such a case, could EU regulators decipher whether an advertisement was AI-generated, or even have the resources to do so? And to what extent could the regulators sufficiently monitor every social media advertisement that lands in the EU?

[xxiii] See supra note 21, art. 50.

[xxiv] See AI Legislation Tracker, American Action Forum,

 https://www.americanactionforum.org/list-of-proposed-ai-bills-table/ (last visited Oct. 6, 2025) for an AI tracker on federal U.S. legislation; see also Orrick, Herrington & Sutcliffe, U.S. AI Law Tracker, Orrick, https://ai-law-center.orrick.com/us-ai-law-tracker/ (last visited Oct. 5, 2025), which tracks AI laws in U.S. states—with California leading the way; Global AI Law and Policy Tracker, iapp, https://iapp.org/resources/article/global-ai-legislation-tracker/ (last visited Oct. 5, 2025), which tracks AI laws globally.      

This is a student blog post and in no way represents the views of the Fordham International Law Journal.