As the use of artificial intelligence ("AI") grows, new disputes arise, especially when it comes to intellectual property rights.
A recent example is the dispute between Studio Ghibli and OpenAI, which began when users started generating anime-style versions of their photos using ChatGPT, mimicking the distinctive Ghibli aesthetic. Currently, a similar problem involves the Midjourney platform, where users are creating images that include characters copyrighted by Disney, Marvel and Universal.
Disney, Marvel and Universal have a long and respected legacy of creating works that have entertained and inspired global audiences for decades. They are behind some of the most iconic, valuable, and enduring copyrighted works across film, television, literature, consumer products, and theme parks. To respond to this unauthorized use, Disney, Marvel, Universal, and other companies (the "Plaintiffs") have filed a lawsuit in California against Midjourney.
I – Legal Grounds for the Lawsuit
The lawsuit is based on two main points: direct infringement and secondary infringement.
1. Direct Copyright Infringement
Subject to the Copyright Act, particularly 17 U.S. Code § 106, copyright owners have the exclusive right to reproduce, publicly display, distribute, and create derivative works based on their protected content.
According to the lawsuit, Midjourney unlawfully infringed on the Plaintiffs' rights mentioned in the article above, in two ways:
- First, during the development and training of its image service;
- Second, in the image outputs it generates for subscribers.
Each unauthorized use is claimed to be a separate and distinct violation. Midjourney is accused of purposely using the Plaintiffs' intellectual property to attract users and profit by offering them countless unauthorized copies and derivative versions of protected works through a three steps' process:
- Collection: Midjourney gathered large volumes of online content using automated tools like bots, scrapers, and web crawlers.
- Cleaning and Reformatting: The collected content was filtered and reformatted into a common technical structure fit for AI training creating new copies in the process.
- Training: The reformatted data, including protected works owned by the Plaintiffs, was used to train the AI model. Although the internal details of the process remain undisclosed, the Plaintiffs allege that the copyrighted content was stored in the system in a way that allows the AI to reproduce, display, and share derivatives of their works without authorization violating their exclusive rights under the Copyright Act.
In comparison, Lebanon, the UAE, and Saudi Arabia also protect copyrights through national laws that give similar rights to what is found in 17 U.S. Code § 106. These countries are signatories of the Berne Convention that sets international minimum standards for copyright protection and applies across all 181 member countries.
In Lebanon, Law No. 75/1999, Article 15, gives authors exclusive rights to control the reproduction, adaptation, distribution, importation, public performance, and communication of their work, protecting against any unauthorized use or modification in any form.
In the UAE, Federal Decree-Law No. 38 of 2021, Article 7, gives authors, their successors, or the copyright holder exclusive rights to reproduce, broadcast, publicly share, modify, lease, and publish the work.
In Saudi Arabia, the Copyright Law issued by Royal Decree No. M/41 in 2003 and amended in 2018, Article 9, grants authors or their designee exclusive rights to print, publish, record, translate, quote, alter, redistribute, communicate to the public, and materially exploit the work, including commercial lease. Financial rights are also granted to performers, composers, producers, and broadcasting organizations as per implementing regulations.
2. Secondary Copyright Infringement
Even if Midjourney claims that its users are the ones creating the infringing content, the Plaintiffs argue that Midjourney is still liable. This is because the platform had full control over the training data and chose to include copyrighted works without authorization, despite having the ability to exclude them.
Additionally, Midjourney has the technical capability to monitor and regulate user prompts and could block prompts likely to generate infringing outputs. The Plaintiffs note that Midjourney already blocks content related to hate speech, nudity, violence, and disrespectful images of public figures demonstrating that the platform has the tools to limit misuse.
II– Can Midjourney Claim "Fair Use"?
The Copyright Act, in section 17 U.S.C. § 107 outlines the fair use doctrine, which permits limited use of copyrighted works without permission for certain purposes such as criticism, commentary, news reporting, teaching, scholarship, or research.
Courts consider the following four factors to determine fair use:
- The purpose and character of the use including whether it is for commercial or nonprofit educational purposes;
- The nature of the copyrighted work;
- The amount and substantiality of the portion used; and
- The effect of the use on the market value of the original work.
While it is ultimately for the court to decide whether Midjourney's actions qualify as fair use, based on the allegations and context, the case appears to be outside the scope of what the fair use doctrine is meant to allow. Midjourney operates a commercial platform with four levels of paid subscriptions ranging between 10$ and 120$ per month, where subscribers can directly generate, reproduce, and publicly display further copies of the copyrighted works. The Plaintiffs' copyrighted works are highly creative, original, and very valuable. It is also alleged that Midjourney used large amounts of these works, possibly even entire pieces, as training data. All these factors strongly weigh against a fair use defense.
It is important to note that neither Lebanon, nor the UAE, or Saudi Arabia apply the "fair use" doctrine as prescribed in Section 107 of the U.S. Copyright Act. Instead, these jurisdictions adhere to the Berne Convention which permits the adoption of limited exceptions to the protection of literary and artistic works under strict conditions, known as the three-step test: such allowed reproduction must (1) occur in certain special cases, (2) not conflict with the normal exploitation of the work, and (3) not unreasonably prejudice the legitimate interests of the author.
Conclusion
This lawsuit could play a key role in shaping the future of AI-generated images. The outcome may set a legal precedent that defines what AI platforms can and cannot do with existing creative content. More importantly, it may finally address the widespread infringement artists and creators are facing in the age of generative AI.
Human art is in real danger due to the misuse of artificial intelligence and this case may be a critical step toward protecting creativity and restoring legal accountability in the digital age.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
