Does Midjourney Steal Art?

Abby Poole

Updated on June 13, 2024

Midjourney is one of the latest players in the burgeoning field of AI art generators. Essentially, it's a text-to-image AI model that can create stunning visuals based on textual descriptions. Imagine typing "a serene sunset over a mountain range" and watching as the AI conjures up a beautiful image that matches your description. Sounds like magic, right? But it's all thanks to advanced machine learning algorithms and vast datasets.

Its capabilities are quite impressive. It can generate images in a wide range of styles, from photo realistic to abstract. It can also create images that are highly detailed and complex, capturing intricate textures and patterns. This makes it incredibly versatile and useful for a variety of applications, from creating marketing materials to designing video game assets.

Moreover, Midjourney is constantly improving. As it is trained on larger and more diverse datasets, their ability to generate high-quality images continues to increase. This means that the images they produce are becoming more realistic and more aligned with the user's descriptions.

However, these tools are not without their limitations. One of the main challenges is that they can sometimes produce images that are not quite what the user had in mind. This is because the AI is interpreting the textual description in its own way, which may not always match the user's vision.

Another limitation is that these tools can sometimes produce images that are derivative or lack originality. This is because they are trained on existing images, which means that they can sometimes replicate elements from those images in their creations. This has led to concerns about whether these tools are "stealing" art from existing artists.

This brings us to the heart of the matter: does Midjourney steal art? The answer is not straightforward. On one hand, Midjourney and other AI art generators are trained on vast datasets of existing images. This means that they are learning from and, to some extent, replicating existing art. This has led to accusations that these tools are infringing on the intellectual property of artists.

On the other hand, proponents of AI art generators argue that these tools are simply using existing images as a reference, much like a human artist might. They argue that the images produced by these tools are original creations, generated by the AI based on the user's description.

Next, let's talk with more detail and deeper about these topics.

The Copyright Controversy

In February 2023, a class-action lawsuit was filed against Midjourney, Stable Diffusion, and other AI image generators by artists like Sarah Andersen, Kelly McKernan, and Karla Ortiz. The lawsuit alleges that these companies have unlawfully used the artists' copyrighted works to train their AI models without permission or compensation. This has sparked a heated debate in the art community and beyond.

One of the most damning pieces of evidence in this case is a spreadsheet allegedly curated by Midjourney. This document lists over 4,700 artists whose distinctive styles the AI can mimic. The implication here is clear: these artists' works were used for training purposes without their consent. Imagine discovering that your unique artistic style has been replicated by an AI, and you didn't even get a heads-up, let alone compensation.

Artists have discovered Midjourney generating images strikingly similar to their unique styles, raising concerns about potential copyright infringement and lack of attribution. It's one thing to be inspired by an artist; it's another to create something that looks like a carbon copy. This has led to a flurry of accusations and a lot of angry tweets.

Critics argue that AI art generators are profiting from human creativity by mechanically blending and replicating artistic styles without transforming the source material in a legally permissible way under fair use. They claim that these AI tools are essentially stealing art and making money off it.

On the flip side, defenders of AI art claim the tools are creating new, original works inspired by existing art, which is a common creative process. They argue that artists have always drawn inspiration from each other, and AI is just another tool in the creative toolbox.

The legal implications are complex, with past court rulings on copyright and transformative use varying across different creative fields like visual art and music. The core issue here is whether training AI models on copyrighted artworks without consent constitutes fair use or infringement.

Is the resulting AI-generated image a truly transformative work or a derivative copy violating intellectual property rights? This question is at the heart of the ongoing legal battle and will likely set a precedent for future cases involving AI and copyright.

Artistic Styles and Potential Infringement

Imagine creating a unique horror/sci-fi illustration style, only to find out that an AI tool like Midjourney can generate images virtually indistinguishable from your work. That's exactly what happened to artist Karla Ortiz. She discovered that when users prompted Midjourney with descriptors matching her style, the results were eerily similar to her own creations.

This raises a critical question: Is Midjourney simply inspired by Ortiz's work, or is it crossing the line into copyright infringement? Critics argue that by training on copyrighted artworks, AI models like Midjourney are effectively copying and replicating the nuanced visual styles of human artists.

In 2022, graphic artist Kristina Kashtanova took legal action against AI art generator Nightcafe. She alleged that the tool had illegally replicated and commercialized her artistic style without her consent. This case highlights the growing concern among artists about the potential exploitation of their intellectual property by AI tools.

Critics argue that AI art generators, by training on copyrighted works, are not just inspired but are effectively copying the unique elements that make an artist's work distinctive. This includes color palettes, textures, and compositions.

Defenders of AI art tools counter that these models are simply inspired by existing artwork, much like human artists are. They argue that AI-generated art falls under transformative fair use, as it creates something new and different from the original works it was trained on.

However, the legal line between permissible inspiration and infringement becomes blurred when AI can so accurately mimic an artist's signature visual elements. The ability of AI to replicate these elements raises complex questions around copyright, creative ownership, and potential exploitation.

The proliferation of AI-generated art could potentially devalue human artists' unique styles. If AI tools can replicate an artist's work without attribution or compensation, it undermines the value of the original creations. This is a significant concern for artists who rely on their distinctive styles to make a living.

Moreover, the lack of clear legal precedents makes it challenging to define the boundaries of what constitutes permissible use of AI in art. As AI technology continues to evolve, these questions will only become more pressing.

Legal Implications and Precedents

Copyright law has always been a bit of a tightrope walk, especially when it comes to transformative use. Courts have had to decide whether a new work adds enough new expression or meaning to be considered "transformative" rather than just a derivative copy.

Take, for example, the case of Campbell v. Acuff-Rose Music, Inc. (1994), where the Supreme Court ruled that 2 Live Crew's parody of Roy Orbison's "Oh, Pretty Woman" was a transformative use. The court emphasized that the new work must add something new, with a further purpose or different character, altering the first with new expression, meaning, or message.

But here's the kicker: these rulings have varied widely across different creative fields like visual art and music. So, applying these precedents directly to AI-generated art is like trying to fit a square peg into a round hole.

The legal landscape is in dire need of evolution to keep pace with AI-generated art. Current copyright laws weren't designed with AI in mind, and this creates a gray area that's ripe for legal battles.

One of the biggest issues is the lack of clear guidelines on using copyrighted works to train AI models. For instance, AI image generators like Midjourney and Stable Diffusion rely on vast datasets, often scraped from the internet, to learn and create. But what happens when these datasets include copyrighted works?

The line between permissible inspiration and outright infringement becomes blurred. AI can mimic an artist's signature style, color palettes, and compositions so accurately that it raises the question: Is this fair use or a derivative copy?

The absence of clear legal precedents specifically addressing the use of copyrighted artworks to train AI models is a significant hurdle. Courts will likely have to interpret existing copyright laws through the lens of this disruptive technology, potentially leading to new legal frameworks tailored to the AI era.

A key question here is whether AI-generated images constitute fair use and transformative works, or if they are derivative copies that violate the original artists' copyrights. The interpretation of what qualifies as a "derivative work" under copyright law is likely to play a crucial role, with different jurisdictions potentially ruling differently.

There's also a growing concern that the proliferation of AI art could devalue human artists' unique styles. Imagine a world where anyone can replicate a famous artist's style without proper attribution or compensation. It's a nightmare scenario for many in the creative community.

Some legal experts argue that new frameworks may be needed to address these unique challenges. Existing copyright laws were not designed with AI in mind, and this technology is pushing the boundaries of what those laws can handle.

So, what are the potential solutions? Here are a few ideas that have been floated around:

Establishing Clear Guidelines: Clearer guidelines on what constitutes fair use and transformative works in the context of AI-generated art.
Redefining Legal Concepts: Redefining legal concepts like "inventorship" to account for AI's role in the creative process.
AI-Specific Patents: Exploring novel classifications like AI-specific patents to protect both human and AI-generated works.

Conclusion

In the end, the question of whether Midjourney steals art is a complex one. It touches on issues of intellectual property, creativity, and the role of AI in the creative process. As these tools continue to evolve, it's likely that this debate will continue. For now, it's clear that AI art generators like Midjourney are powerful tools that offer new possibilities for creating and exploring visual art. Whether they are "stealing" art or simply drawing inspiration from existing works is a question that each user will have to answer for themselves.

I hope that you have learned something from this article and they can help you find the answer to the question, i.e. "Does Midjourney steal art?". So what is your opinion now?

Resources:

Make Your 512p/1024p AI Arts Bigger and Clearer

Does Midjourney Steal Art?

The Copyright Controversy

Artistic Styles and Potential Infringement

Legal Implications and Precedents

Conclusion

You May Also Like