AI-Generated Images on Your Business Website (EU 2026)

You have been using DALL-E to make blog header images for a year. A new client asks whether you need a licence to publish them. The short answer for most ordinary marketing images is no. The four risk layers below explain when that answer changes and what the AI Act's Article 50(4) adds from 2 August 2026.

<figure className="my-8"> <svg role="img" aria-labelledby="ai-image-tree-title" aria-describedby="ai-image-tree-desc" viewBox="0 0 1200 900" xmlns="http://www.w3.org/2000/svg" style={{ maxWidth: '100%', height: 'auto' }}> <title id="ai-image-tree-title">Four-layer decision tree for using AI-generated images on a business website.</title> <desc id="ai-image-tree-desc">Top-down decision tree starting with the question should I publish this AI-generated image. Four sequential layers test for reproduction of a specific existing work, presence of trademarks or watermarks, identifiable real people and resemblance to authentic depictions of real persons or events. The first two layers, if answered yes, lead to red stop nodes. Layers three and four lead to amber caution nodes requiring consent or AI-generated labelling. A no answer at all four layers leads to a green publish outcome with no AI Act labelling obligation.</desc> <rect x="0" y="0" width="1200" height="900" fill="#FFFFFF"/> <rect x="350" y="20" width="500" height="70" rx="10" fill="#1A1A1A"/> <text x="600" y="52" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="18" fontWeight="600" fill="#FFFFFF">I want to publish this AI-generated image</text> <text x="600" y="75" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="13" fill="#E5E5E5">on my business website</text> <path d="M 600 90 L 600 120" stroke="#525252" strokeWidth="2"/> <polygon points="600,125 595,115 605,115" fill="#525252"/> <rect x="270" y="130" width="660" height="80" rx="8" fill="#FFFFFF" stroke="#1A1A1A" strokeWidth="2"/> <text x="600" y="158" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="16" fontWeight="600" fill="#1A1A1A">Layer 1. Does it reproduce a specific recognisable existing work?</text> <text x="600" y="180" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fill="#525252">(specific artwork, photograph, branded character, copyrighted illustration)</text> <text x="600" y="198" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#525252">Common with "in the style of [named living artist]" prompts</text> <path d="M 270 170 L 130 240" stroke="#B91C1C" strokeWidth="2"/> <polygon points="130,245 130,232 142,238" fill="#B91C1C"/> <text x="180" y="210" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="500" fill="#B91C1C">Yes</text> <path d="M 600 210 L 600 248" stroke="#1B7D56" strokeWidth="2"/> <polygon points="600,253 595,243 605,243" fill="#1B7D56"/> <text x="615" y="235" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="500" fill="#1B7D56">No</text> <rect x="40" y="250" width="200" height="65" rx="8" fill="#B91C1C"/> <text x="140" y="278" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="600" fill="#FFFFFF">STOP. Copyright risk.</text> <text x="140" y="298" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#FFFFFF">Replace the image.</text> <rect x="270" y="260" width="660" height="80" rx="8" fill="#FFFFFF" stroke="#1A1A1A" strokeWidth="2"/> <text x="600" y="288" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="16" fontWeight="600" fill="#1A1A1A">Layer 2. Does it contain a trademark, logo or watermark?</text> <text x="600" y="310" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fill="#525252">(brand logos, the Getty or iStock watermark, recognisable mascots)</text> <text x="600" y="328" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#525252">The point on which Stability AI was found partially liable</text> <path d="M 270 300 L 130 370" stroke="#B91C1C" strokeWidth="2"/> <polygon points="130,375 130,362 142,368" fill="#B91C1C"/> <text x="180" y="340" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="500" fill="#B91C1C">Yes</text> <path d="M 600 340 L 600 378" stroke="#1B7D56" strokeWidth="2"/> <polygon points="600,383 595,373 605,373" fill="#1B7D56"/> <text x="615" y="365" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="500" fill="#1B7D56">No</text> <rect x="40" y="380" width="200" height="65" rx="8" fill="#B91C1C"/> <text x="140" y="408" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="600" fill="#FFFFFF">STOP. Trademark risk.</text> <text x="140" y="428" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#FFFFFF">Replace the image.</text> <rect x="270" y="390" width="660" height="80" rx="8" fill="#FFFFFF" stroke="#1A1A1A" strokeWidth="2"/> <text x="600" y="418" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="16" fontWeight="600" fill="#1A1A1A">Layer 3. Does it contain identifiable real people?</text> <text x="600" y="440" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fill="#525252">(faces of named individuals, celebrities, employees, customers)</text> <text x="600" y="458" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#525252">GDPR Art. 9 + national personality rights</text> <path d="M 270 430 L 130 500" stroke="#D97706" strokeWidth="2"/> <polygon points="130,505 130,492 142,498" fill="#D97706"/> <text x="180" y="470" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="500" fill="#D97706">Yes</text> <path d="M 600 470 L 600 508" stroke="#1B7D56" strokeWidth="2"/> <polygon points="600,513 595,503 605,503" fill="#1B7D56"/> <text x="615" y="495" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="500" fill="#1B7D56">No</text> <rect x="40" y="510" width="200" height="80" rx="8" fill="#D97706"/> <text x="140" y="540" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="600" fill="#FFFFFF">PAUSE.</text> <text x="140" y="558" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#FFFFFF">Get consent or use</text> <text x="140" y="572" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#FFFFFF">a clearly fictional</text> <text x="140" y="586" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#FFFFFF">framing.</text> <rect x="270" y="520" width="660" height="90" rx="8" fill="#FFFFFF" stroke="#1A1A1A" strokeWidth="2"/> <text x="600" y="548" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="16" fontWeight="600" fill="#1A1A1A">Layer 4. Would it appear authentic to a reasonable viewer?</text> <text x="600" y="570" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fill="#525252">(a real-looking depiction of a real person, place or event)</text> <text x="600" y="588" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#525252">Deepfake definition in Article 3(60) AI Act</text> <path d="M 270 570 L 130 640" stroke="#D97706" strokeWidth="2"/> <polygon points="130,645 130,632 142,638" fill="#D97706"/> <text x="180" y="610" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="500" fill="#D97706">Yes</text> <path d="M 600 610 L 600 648" stroke="#1B7D56" strokeWidth="2"/> <polygon points="600,653 595,643 605,643" fill="#1B7D56"/> <text x="615" y="635" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="500" fill="#1B7D56">No</text> <rect x="40" y="650" width="200" height="80" rx="8" fill="#D97706"/> <text x="140" y="680" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="13" fontWeight="600" fill="#FFFFFF">LABEL.</text> <text x="140" y="698" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#FFFFFF">Mark as AI-generated</text> <text x="140" y="712" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#FFFFFF">under Art. 50(4) AI Act</text> <text x="140" y="726" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#FFFFFF">from 2 Aug 2026.</text> <rect x="350" y="680" width="500" height="100" rx="10" fill="#1B7D56"/> <text x="600" y="715" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="18" fontWeight="600" fill="#FFFFFF">PUBLISH.</text> <text x="600" y="740" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="13" fill="#FFFFFF">Standard marketing image. No AI Act labelling obligation.</text> <text x="600" y="760" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fill="#E5E5E5">Note. You probably do not own the copyright. Anyone may re-use it.</text> <text x="600" y="820" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#525252">Source: Regulation (EU) 2024/1689, Articles 3(60), 50. Getty Images v Stability AI [2025] EWHC 2863 (Ch).</text> </svg> <figcaption>The four questions to ask before publishing. Most generic marketing images end at the green "publish" node.</figcaption> </figure>

The short answer

Generic AI marketing imagery is usually fine to publish. The risk concentrates in four places: images that reproduce a specific existing work, images that carry a brand logo or stock-photo watermark, images that contain identifiable real people and images that look authentic enough to pass for real photography. If you have ever wondered whether the kind of Getty demand letter people receive for stock-photo issues could land for an AI image, it can. But only when the output reproduces a recognisable existing photograph or contains a Getty watermark. The rest of the time, the worst that happens is you find out you do not own the image you "made."

What the courts have actually decided

The most-cited ruling is Getty Images (US) Inc & Ors v Stability AI Ltd [2025] EWHC 2863 (Ch), a UK High Court judgment of 4 November 2025. You can read the full judgment on the Judiciary website. The popular headline was "Stability AI mostly wins." That is roughly true, but the holdings are narrower than the headline suggests and EU readers should treat it as UK-specific context, not as a settled EU position.

Getty abandoned its primary copyright infringement claim mid-trial because it could not prove the relevant training activity took place in the UK. The court did not rule on whether training a diffusion model on copyrighted images is itself an infringement. The "the court said training is fine" reading of the judgment is wrong. The court did not reach the question.

The court did rule on secondary copyright infringement under sections 22, 23 and 27 of the Copyright, Designs and Patents Act 1988. It held that Stable Diffusion model weights are not a "copy" of the training images. The model contains trained parameters, not stored reproductions. Trademark infringement under sections 10(1) and 10(2) of the Trade Marks Act 1994 partially succeeded on a narrow set of synthetic images that bore the Getty or iStock watermark. The section 10(3) claim and passing-off were rejected. Getty has indicated it will use the findings of fact in its parallel US litigation.

EU-side litigation is unsettled. Robert Kneschke v LAION in Germany produced a 2024 first-instance decision favourable to LAION on a TDM exception, with an appeal expected. Like Company v Google Ireland Limited (Case C-250/25) is a CJEU referral pending. The EDPB's December 2024 Opinion 28/2024 on AI models addressed GDPR aspects of training but did not resolve copyright. Treat the EU position as open for now.

The four risk layers for your business

Layer 1. Does it reproduce a specific recognisable existing work?

Most AI outputs do not. Some do. Diffusion models occasionally produce near-copies of training images, and prompts that name a living artist, a recent film or a copyrighted character raise that risk substantially. A prompt for "a friendly dentist's office" is low-risk. A prompt asking for "in the style of [named living photographer]" is not. A reverse image search on hero images and paid-ad visuals will usually surface any close original within a minute.

Layer 2. Does it contain a trademark, logo or watermark?

This is the layer the Getty ruling actually touched. Stable Diffusion was generating images with the Getty and iStock watermarks visible, and the UK court found that this constituted trademark infringement on those specific outputs. The same logic applies if your AI image contains a recognisable brand logo, a sports team crest, a film studio mascot or a stock-agency watermark fragment.

Watermarks are the easiest defect to catch. Zoom in to 200% on the corners of any hero image before publishing. If the firms that handle these claims commercially in the EU send you a letter, how Copytrack and PicRights demands work in practice covers what to expect.

Layer 3. Does it contain identifiable real people?

A photograph of an identifiable real person is personal data under GDPR Article 4(1). An AI image that resembles an identifiable real person raises the same controller-level obligations under Article 5 and may engage special-category processing under Article 9 if the image conveys sensitive attributes. Member State law adds national personality rights on top, which vary in strength across the EU.

The practical rule is the same as for ordinary photography. If you want to publish an image that resembles a specific real person, you need their consent. AI does not change the rule. If you have no consent, prompt explicitly for "a fictional person, not resembling any real individual" and check the output before publishing.

Layer 4. Do you "own" what you made?

Probably not, in any useful sense. The US Copyright Office stated in March 2023 and reaffirmed in 2024 that pure AI-generated output without sufficient human creative input is not eligible for copyright. The EU's Infopaq test (Case C-5/08) requires "the author's own intellectual creation," which a prompt-only output is unlikely to satisfy. National courts have not produced a clear ruling but are likely to track the same direction.

That ownership gap matters less than people fear. Most SMBs use marketing images to communicate, not to license out. The practical implication is that any competitor can re-use the same image you generated. It is a marketing concern. The broader question of who pays when AI-related compliance goes wrong on a website is more layered, and the answer is rarely "the AI vendor."

Article 50 of the AI Act, applicable 2 August 2026

The EU AI Act applies from 2 August 2026 in relation to its Article 50 transparency obligations. Two paragraphs of Article 50 matter for images on a website.

Article 50(2) puts a marking obligation on the provider of a generative AI system. Midjourney, OpenAI, Stability and the rest must mark their outputs as artificially generated in a machine-readable way. This is not your obligation as the website operator. Your only duty is not to intentionally strip vendor watermarks.

Article 50(4) puts a labelling obligation on the deployer (you) when the AI-generated content is a deepfake under Article 3(60). The definition is narrow.

<figure className="my-8"> <svg role="img" aria-labelledby="deepfake-matrix-title" aria-describedby="deepfake-matrix-desc" viewBox="0 0 900 540" xmlns="http://www.w3.org/2000/svg" style={{ maxWidth: '100%', height: 'auto' }}> <title id="deepfake-matrix-title">Two-by-two matrix for whether an AI-generated image is a deepfake under Article 50(4) of the AI Act.</title> <desc id="deepfake-matrix-desc">Matrix with two axes. The horizontal axis asks whether the image resembles a real subject. The vertical axis asks whether the image would appear authentic to a reasonable viewer. The top-right quadrant where both answers are yes is a deepfake under Article 3(60) and requires labelling from 2 August 2026. The top-left quadrant is an edge case warranting documented editorial reasoning. The bottom-right and bottom-left quadrants fall outside Article 50(4).</desc> <rect x="0" y="0" width="900" height="540" fill="#FFFFFF"/> <text x="450" y="34" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="18" fontWeight="600" fill="#1A1A1A">Is this AI image a deepfake under Article 50(4)?</text> <text x="450" y="58" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fill="#525252">Both conditions must be present.</text> <text x="120" y="95" fontFamily="DM Sans, sans-serif" fontSize="12" fontWeight="500" fill="#525252">Would appear authentic</text> <text x="120" y="111" fontFamily="DM Sans, sans-serif" fontSize="12" fontWeight="500" fill="#525252">to a reasonable viewer</text> <text x="155" y="155" fontFamily="DM Sans, sans-serif" fontSize="12" fontWeight="600" fill="#1A1A1A">Yes</text> <text x="155" y="345" fontFamily="DM Sans, sans-serif" fontSize="12" fontWeight="600" fill="#1A1A1A">No</text> <text x="450" y="500" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fontWeight="500" fill="#525252">Resembles a real subject</text> <text x="290" y="478" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fontWeight="600" fill="#1A1A1A">No</text> <text x="650" y="478" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="12" fontWeight="600" fill="#1A1A1A">Yes</text> <rect x="200" y="130" width="240" height="130" rx="8" fill="#FEF3C7" stroke="#D97706" strokeWidth="2"/> <text x="320" y="160" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="14" fontWeight="600" fill="#D97706">Edge case</text> <text x="320" y="186" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Generic but uncannily real.</text> <text x="320" y="204" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Probably outside Art. 50(4).</text> <text x="320" y="222" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Document your editorial</text> <text x="320" y="238" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">reasoning if asked.</text> <rect x="460" y="130" width="240" height="130" rx="8" fill="#FEE2E2" stroke="#B91C1C" strokeWidth="2"/> <text x="580" y="160" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="14" fontWeight="600" fill="#B91C1C">Deepfake</text> <text x="580" y="186" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Caught by Article 3(60).</text> <text x="580" y="204" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Label as AI-generated</text> <text x="580" y="222" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">from 2 August 2026.</text> <text x="580" y="240" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Up to EUR 15M or 3% turnover.</text> <rect x="200" y="280" width="240" height="130" rx="8" fill="#DCFCE7" stroke="#1B7D56" strokeWidth="2"/> <text x="320" y="310" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="14" fontWeight="600" fill="#145E40">Outside scope</text> <text x="320" y="336" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Pure invented imagery.</text> <text x="320" y="354" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">No real subject, no</text> <text x="320" y="372" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">authentic feel.</text> <text x="320" y="390" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Publish without label.</text> <rect x="460" y="280" width="240" height="130" rx="8" fill="#DCFCE7" stroke="#1B7D56" strokeWidth="2"/> <text x="580" y="310" textAnchor="middle" fontFamily="Instrument Serif, serif" fontSize="14" fontWeight="600" fill="#145E40">Outside scope</text> <text x="580" y="336" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">Recognisable but obviously</text> <text x="580" y="354" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">stylised (cartoon, sketch).</text> <text x="580" y="372" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">No risk of being mistaken</text> <text x="580" y="390" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="11" fill="#1A1A1A">for an authentic depiction.</text> <text x="450" y="535" textAnchor="middle" fontFamily="DM Sans, sans-serif" fontSize="10" fill="#525252">Article 50(4) labelling applies only to deepfakes under Art. 3(60) or AI text on public-interest matters without editorial review.</text> </svg> <figcaption>The narrower-than-people-think scope of AI Act image labelling. Both axes need to read "yes" before the labelling obligation engages.</figcaption> </figure>

A real estate agent generating "virtual staging" of an actual listing that could pass for a real photograph of that specific room sits in the red quadrant and needs an "AI-generated" label. A bakery using AI to depict a generic croissant on the menu page does not. The maximum fine under Article 99 for an Article 50 breach is the higher of EUR 15 million or 3% of worldwide annual turnover, but Article 99(6) tells Member States to apply lower fines to SMEs, and realistic enforcement against a small business is far below the maxima. The full AI Act guide for website owners covers the other paragraphs of Article 50 as well as chatbot disclosure.

A 7 May 2026 provisional Digital Omnibus on AI may give providers of generative AI systems already on the market a transitional period until 2 December 2026 for Article 50(2) marking. The 2 August 2026 application of Article 50 itself does not move. Verify status before acting.

Practical rules for your business

A short checklist for any EU SMB using AI imagery on a website.

1. Do not prompt with "in the style of [named living artist]." It elevates copyright risk without much creative benefit.
2. Zoom in to 200% on every AI image before publishing and look for stray watermarks, brand logos and signature fragments in the corners.
3. If you need an image of a person, prompt for "a fictional person, not resembling any real individual." If you want a real person, get their written consent. AI does not change that rule.
4. Keep prompt logs. A folder of "prompt + date + tool" records is cheap insurance if a question is ever raised.
5. If you publish an image that depicts a real-looking person, place or event, label it as AI-generated. A short caption is enough and keeps you safe under Article 50(4).
6. Do not claim copyright on AI-generated images in marketing contracts. You may not have any to assign.
7. For high-stakes images (campaign hero, packaging, legal documents), use a licensed library or a human illustrator. Our recommendations on free stock photo sources that are safe to use are a starting point.

If you want a check on the rest of the site at the same time, our free compliance scan covers GDPR, cookies, accessibility and image rights. It does not yet check AI-image legitimacy directly. AI-image checks are on the roadmap. The scan will tell you whether the surrounding site is on solid ground.

What this article does not tell you

It does not tell you whether the AI provider was entitled to use the training data. That question stays open in the EU and is contested in Germany, in California and at the CJEU. It does not tell you that an AI image is automatically safe. The four layers can each go wrong. And it does not tell you what happens if your AI provider is eventually found liable for its training. The cascade through their terms of service is untested.

The cluster pieces this article connects to cover the related strands. Who is liable when AI helps build a website sits next to this one. The same agency-tool-operator chain applies whether the AI generated code or generated an image.

Common Questions

Can someone send me a Getty-style demand letter for an AI-generated image?

Possible, but only when the AI output reproduces a recognisable existing photograph or contains a Getty or iStock watermark. Getty v Stability AI confirmed that point in November 2025. For typical AI marketing imagery with no watermarks and no recognisable real subjects, the risk of a demand letter stays low.

Do I own the AI image I generated for my website?

Probably not. The EU's Infopaq test requires the author's own intellectual creation. Pure prompt-based output usually fails it. A competitor can re-use the same image, and that is a marketing concern more than a legal one. The labelling question and the ownership question are separate.

Do I have to label AI-generated images on my website under the AI Act?

Only when the image is a deepfake under Article 3(60), meaning it resembles a real person, place or event and would appear authentic to a reasonable viewer. Article 50(4) labelling applies from 2 August 2026. Generic AI marketing imagery without identifiable real subjects falls outside.

Did Getty v Stability AI settle the training-data question for the EU?

No. Getty abandoned its primary copyright infringement claim mid-trial because the training did not happen in the UK. The court ruled only on secondary copyright and trademark. EU-side litigation includes Robert Kneschke v LAION in Germany and Case C-250/25 at the CJEU. The picture stays unsettled.

Is Midjourney safer than DALL-E or Stable Diffusion for business use?

No meaningful legal difference today. All three rely on similar training-data practices and their terms of service all push output risk back to the user. Pick on quality and licensing terms, not on a notional safety advantage.

Related reading

The image cluster on this site sits around four pieces:

• Received a Getty Images letter. The crisis guide for human-sourced image claims.
• Are Copytrack and PicRights demands legitimate. Companion crisis article.
• Free stock photo sources. Licensed alternatives that stay clear of demand-letter exposure.
• Web designer copyright liability under EU law. When the issue was introduced by your agency.

The AI cluster pieces this article connects to:

• AI-built website liability under EU law. The hub article on GDPR, EAA and cookie liability when AI helped build the site.
• The EU AI Act for website owners. Article 50 in depth, including chatbot disclosure and the editorial-review exemption for AI-drafted text.

This article is technical analysis, not legal advice. The author is not your lawyer and is not your registered controller. For a binding view, talk to one of those.