Midjourney and the EU AI Act: Who Must Disclose AI Images, When Exemptions Apply, and How to Comply

Do Midjourney images require disclosure under the EU AI Act? The short answer: yes for most commercial applications, but the compliance landscape is more nuanced than many creators realize.

Article 50 of the EU AI Act, which applies from 2 August 2026, requires deployers of AI systems that generate synthetic content to clearly disclose when images, audio, or video have been artificially created. This directly affects anyone using Midjourney for business purposes, marketing campaigns, e-commerce, or other public-facing content within the EU market.

The key distinction isn’t just that you’re using Midjourney, but how you use it, who you target, and which exemptions might apply. While Midjourney Inc. has certain provider obligations, the primary disclosure responsibility falls on you as the end user when deploying these images commercially.

Understanding this framework is essential for creators, marketers, and business owners looking to leverage Midjourney’s capabilities without running afoul of EU regulations. The stakes include potential penalties and compliance audits that could significantly impact your operations.

Article 50 imposes transparency obligations on "artificial or manipulated image, audio or video content," as defined by the Code of Practice on Transparency of AI-Generated Content. The regulation targets content that "appreciably resembles existing persons, objects, places or other entities or events and would falsely appear to a person to be authentic or truthful."

For a deployer, the disclosure duty under Article 50(4) triggers only when the image is a deep fake — it resembles real persons, objects, places or events and could be taken as authentic (Article 3(60)). Photorealistic Midjourney images in marketing often meet this because they convincingly depict realistic products, people or scenes; stylised or obviously-unreal images may not. Separately, Midjourney as provider must mark its outputs as AI-generated under Article 50(2), whatever the style.

Deployers must implement "appropriate and clear" disclosure measures—not just fine print, but prominent labels that immediately signal the content's artificial nature. As we explain in our comprehensive guide to Article 50, the regulation emphasizes transparency over restriction to preserve public trust without stifling AI creativity.

Importantly, this obligation applies regardless of an AI system's high-risk classification under other AI Act provisions. Article 50 establishes a separate, broader transparency requirement that covers most commercial uses of tools like Midjourney when generating content for public consumption.

The EU AI Act defines distinct obligations for each stakeholder in the AI deployment chain; understanding your role determines your specific compliance requirements.

Midjourney Inc. as AI System Provider must build transparency capabilities into its platform and provide tools for proper disclosure. This includes features to identify and label AI-generated content, maintaining system documentation, and supporting downstream compliance.

You as the Deployer/End User bear the primary disclosure obligation when using Midjourney images commercially. Whether you're a marketing agency, e-commerce retailer, or content creator, you're responsible for applying Article 50's "appropriate and clear" labels.

The provider vs deployer obligations framework provides a clear responsibility matrix:

When multiple parties are involved—such as agencies creating content for clients—they may share disclosure responsibilities. The deploying agency assumes immediate compliance obligations, while the client publishing the content must ensure labels reach end users. Clear contractual agreements should define who handles disclosure implementation and ongoing compliance monitoring.

For Guidelines on transparency obligations for providers and ..., the regulatory framework emphasizes that responsibility follows control over content deployment rather than content creation alone.

For deep-fake images, Article 50(4) provides one attenuated treatment rather than a full exemption: where the image is part of an evidently artistic, creative, satirical or fictional work, disclosure is limited to an appropriate acknowledgement that does not spoil the work. There is no separate "editorial" exemption for images — the editorial-responsibility exemption in Article 50(4) applies to text, not images.

Artistic Expression Exemption covers works where the AI-generated image is itself the artistic expression, rather than serving commercial purposes. A digital artist selling explicitly AI art may qualify, but a furniture retailer using Midjourney for product lifestyle images typically would not. Qualification hinges on whether the AI generation is the communicated artistic message.

Satirical Content Boundaries protect parody, commentary, and similar creative expression. Political cartoonists using Midjourney for satirical imagery may claim this exemption, but businesses using AI-generated humor for marketing likely cannot. It requires clear labeling of satirical intent and sufficient editorial context to make the commentary purpose evident.

Journalistic/editorial use has no image exemption of its own under Article 50 — a photorealistic AI news image is still a deep fake requiring disclosure (the guidelines note journalistic images demand more care, not less). News organizations creating illustrations for complex stories may qualify, provided the images are clearly marked as illustrative rather than documentary. This exemption also demands adherence to journalism standards and robust editorial oversight.

Our detailed breakdown of disclosure exceptions emphasizes that qualifying for exemptions requires proactive documentation. You must demonstrate that your use case genuinely fits an exemption and maintain audit trails showing consistent application of the criteria across your content processes.

The key principle: exemptions protect expressive and editorial freedom but do not shield commercial deception or marketing that could mislead consumers about the nature of the content they're viewing.

Different Midjourney applications trigger different Article 50 compliance requirements, so scenario-specific guidance is essential.

E-commerce Product Images represent the highest-risk category. Using Midjourney to create lifestyle shots, product mockups, or promotional imagery requires clear AI disclosure because these images influence purchase decisions. Disclosure must appear prominently—as a visible text overlay, icon placement, or immediate caption. Photorealistic product imagery is exactly what the deep-fake test captures, since consumers expect product images to be authentic unless clearly marked otherwise.

Social Media Advertising demands platform-specific disclosure. Instagram posts using Midjourney images need visible AI labels within the image or prominent hashtags like #AIGenerated. Facebook advertising policies may require disclosure in ad copy. LinkedIn business content should include AI disclosure in post text. Platform-specific guidance from What Are the Platform-Specific Rules We Need to Follow? shows how disclosure requirements interact with existing platform policies.

Editorial and News Content may qualify for exemptions when used as illustrations, but news organizations must uphold clear editorial standards. There is no dedicated editorial exemption for images; but where a conceptual graphic is clearly stylised — not a realistic depiction that could be mistaken for a real photo — it may fall outside the deep-fake definition, whereas a photorealistic news image does not. However, using AI to generate fake documentary-style images violates both disclosure requirements and journalism ethics.

Personal vs Commercial Use Boundaries create key distinctions. Personal social media posts using Midjourney generally don’t trigger disclosure unless monetized or promotional. Influencer content, business profiles, or any commercial promotion requires disclosure regardless of account type.

B2B Marketing Materials need disclosure when targeting EU markets, including presentations, white papers, case studies, and promotional content. Print materials require embedded disclosures that travel with the image, while digital materials can use metadata-based solutions alongside visible labeling.

The compliance matrix shows that commercial intent, public distribution, and potential consumer impact determine disclosure necessity more than technical AI usage alone.

Effective compliance demands disclosure methods that satisfy legal requirements and usability across various content distribution channels.

EU-Recommended Disclosure Formats include standardized icons and text labels for cross-platform recognition. The Commission suggests visible watermarks, text overlays stating "AI Generated," or standardized symbols that clearly signal artificial origin. These indicators must be clear and distinguishable (Article 50(5)), noticeable by an average viewer without detailed inspection.

C2PA Content Credentials Implementation offers technical-grade disclosure via metadata that travels with images across platforms. The Provenance and Authenticity Standards Assessment Working ... demonstrates how blockchain-based provenance tracking ensures verifiable AI disclosure persists through image sharing and platform transfers. Midjourney users can implement C2PA via compatible editing software and distribution platforms.

Platform-Specific Disclosure Methods require tailored approaches:

Our AI artwork disclosure best practices outlines creative implementation approaches that preserve aesthetic quality while ensuring compliance.

Technical Implementation Approaches include automated watermarking, API-based labeling workflows, and CMS integration. The goal is to create disclosure systems that scale with content production while ensuring consistency and audit trail documentation.

Audit Trail Creation involves documenting disclosures across all channels, preserving evidence of compliance, and establishing review processes to catch gaps before public release.

Article 50 obligations apply across EU member states from 2 August 2026, with enforcement mechanisms varying by jurisdiction but following the AI Act’s common penalty framework.

Implementation Timeline: the AI Act entered into force in August 2024, and its Article 50 transparency obligations apply from 2 August 2026. Organizations using Midjourney commercially should have disclosure systems in place before that date.

Penalty Structure follows the AI Act’s tiered approach, with Article 50 violations potentially incurring fines up to €15 million or 3 % of worldwide annual turnover, whichever is higher. These maximum penalties target serious violations such as systematic deception, but even minor lapses can prompt audits and significant operational costs.

Enforcement Mechanisms include market surveillance, consumer protection, and data protection authorities, depending on member-state implementation. They will likely prioritize high-visibility cases: consumer deception, non-disclosure for competitive gain, and systematic violations by large operators.

Grace Periods and Implementation Support may vary by member state, with some authorities focusing on education and guidance during initial enforcement phases. However, organizations should not rely on forbearance and must deploy compliant disclosure systems as soon as practical after development and testing.

Proactive compliance significantly reduces regulatory risk compared to reactive measures once enforcement begins.

Achieving Midjourney compliance under the EU AI Act requires identifying your role as a deployer, applying tailored disclosure methods for your use cases, and maintaining documentation to demonstrate compliance. Although artistic and editorial uses may be exempt, most commercial applications require clear AI labels that accompany content across all distribution channels.

The 2 August 2026 application date gives you time to prepare, but the technical and procedural requirements mean you should begin building your disclosure system now. For step-by-step guidance, consult our complete Article 50 compliance checklist to ensure your Midjourney workflows meet EU requirements while preserving creative and commercial flexibility.

This article is general information, not legal advice.