Can AI Be Copyrighted?
The rise of artificial intelligence (AI) has had a profound impact on many industries, including the arts. Artists around the world are now creating AI-generated art, and it is raising several copyright questions.
I generated the image above using Adobe's Firefly, and this section shows the resolution increased 6x, which will allow it to be reproduced at wall size. There are many others, but I have also used Midjourney, which requires a Discord subscription. Then your subscription gives you an allotment of time with the Graphics Processing Units. Once your allotted time is up, you’re out of luck.
How Is AI Art Generated?
Like many types of AI, AI art generators use various algorithms and datasets to obtain, organize, and reproduce information. There are different types of generators, such as the standard text-to-image generator, whereby you can type a description, and the AI tool will generate art to match. You can also directly feed in your own images to create new AI pieces.
The algorithms use information from existing art pieces, styles, and imagery. AI-generated art is not newly imagined but made from existing ideas and images. This could be problematic since it often happens without the original artists’ consent.
Throughout 2022, the AI art generation became readily available to the public, and, as such, anyone with access to the internet can create and share AI artwork of their own. What does this mean for copyright around AI art?
Can AI Art Be Copyrighted?
Since AI art is created by algorithms, computers, and cross-wired information gathered over time, there is no one artist of a single AI art piece. By that logic, an AI art piece cannot be copyrighted by typical copyrighting standard practices.
One AI art generator, Deep AI, states on its terms of service page that all content created using its AI tools is copyright-free, including for all legal uses, such as personal and commercial gain. Adobe believes that its AI ethics allow it to differentiate its generative AI offerings from those of its competition.
In the United States, copyright authorship can only be granted to works created by a human that is sufficiently original, as well as a short list of other requirements. According to Shanti Escalante-De Mattei, writing in Artnews, “copyright only protects creations made by humans, [and that] will be the guiding principle for future judgments of the registration of works. When evaluating a work submitted for registration, copyright officials will be tasked with judging if the original choices executed in a work were produced by a human mind or produced mechanically.”
Of course, AI-produced art is not authored by humans directly nor made of original materials. Most countries worldwide follow similar practices, making AI-generated artwork unable to be copyrighted.
In the United States, copyright law protects "original works of authorship" fixed in a tangible medium of expression. This means that AI-generated art can be copyrighted by the “prompt engineer” who directs the AI application, but only if it meets the requirements of originality and fixation.
This is a relatively low bar to meet. A work is considered original if it is the product of the author's (human’s) own creative thought. This means that AI-generated art can be copyrighted even if created using a pre-existing algorithm or dataset.
This requires that a work be recorded in a tangible medium of expression. This can be done in various ways, including writing, recording, photographing, drawing, or painting. AI-generated art can be fixed in multiple ways, including saving it as a digital file or printing it on paper.
Classic and Contemporary Artists
AI art is not a new phenomenon. In the early 1960s, computer scientist and artist Frieder Nake created a series of computer-generated images exhibited in galleries worldwide. In the 1970s, artists such as Harold Cohen and Charles Csuri began to use computers to create more complex and sophisticated works of art.
In recent years, there has been a resurgence of interest in AI art, as new technologies have made it possible to create more realistic and visually appealing images. Some of the most well-known contemporary AI artists include Robert Reed, Mario Klingemann, and Robbie Barrat.
There are a few rights applicable to AI works generated by original human thought, as discussed in Part One of this series. However, unique challenges remain to be addressed through regulations and statutes. In the meantime, the design community and AI-generation platforms will work at self-regulation and attempt to safeguard their generated content.
Once an AI-generated work is copyrighted, the author has the exclusive right to reproduce, distribute, perform, display, and create derivative works. This means the author can prevent others from copying, selling, or displaying the work without their permission.
However, there are some exceptions to copyright protection. For example, fair use allows for the limited use of copyrighted material without permission, including demonstrating (as in teaching), criticism, commentary, or news reporting.
The copyright status of AI-generated art is still evolving, and a few questions need to be answered. For example, it is still being determined whether AI-generated art can be copyrighted if created using a pre-trained algorithm or dataset. Additionally, it needs to be clear how copyright law will apply to AI-generated art developed collaboratively between humans and AI.
As AI technology continues to develop, it is clear that copyright law will need to be updated to address the unique challenges posed by AI-generated art.
In February 2023, Getty Images opened a lawsuit against Stability AI, an AI generator Getty suspects of misusing more than 12 million Getty photos to create AI images. London-based Stability AI released Stable Diffusion, an AI-based system for generating images from text inputs, and image generator DreamStudio last August. So, while the resulting images hold no copyright, Getty claims the photos used to create them had copyright licenses that weren’t adhered to.
To avoid problems, consider how the generator you’re using may have obtained its information. Different AI generators will have different rules for their use, and since AI art generation is such a new field, you should read the terms and conditions before using or distributing any artwork you create through AI.
Getty and Adobe
Two of the most prominent image licensing companies, Getty Images and Adobe have recently taken steps to address the issue of copyright in AI art. In 2022, Getty Images announced it would no longer accept AI-generated images for licensing. However, Adobe has developed a new tool that allows users to identify AI-generated images.
Adobe is taking several steps to manage the use of its Firefly AI-generated images. While still in beta, the Firefly images can be helpful in wide-format décor applications. My experiments increasing the 1.2 Megabyte images to 100 Megabytes produced outstanding results.
Adobe says the first model of Firefly is trained to be commercially safe and is initially focused on generating images and text effects. It was trained on a massive dataset that includes images from various sources, including Adobe Stock, Wikimedia Commons, and the public domain. This helps to ensure that the images generated by Firefly are of the highest quality, free from copyright infringement, and, therefore, safe for commercial use.
The AI ethics principles that guide the development and use of Firefly include accountability, responsibility, and transparency, helping to ensure that Firefly is used responsibly and ethically.
By providing users with tools to control the use of Firefly mages, users can watermark their images or restrict their use to personal or non-commercial purposes. This helps to ensure that Firefly mages are used in a way that is respectful of the rights of others.
Adobe is committed to being accountable for the use of Firefly. This means that Adobe will be transparent about how Firefly works and will take steps to mitigate any potential risks.
In its commitment to responsibility, Adobe will use Firefly in a way that benefits society and respects all users' rights. It is working with partners such as the Partnership on AI to educate the public about the benefits and risks of AI-generated content. This includes providing information about how AI-generated content can be used for good and how it can be used for bad.
Regarding transparency, Adobe is going to be transparent about how Firefly works. Adobe will publish information about Firefly, including its training data, algorithms, and results.
These moves by Getty Images and Adobe suggest that the copyright landscape for AI art is still evolving. It remains to be seen how copyright law will ultimately be applied to this new and emerging art form.
The copyright status of AI art is a complex and evolving issue. As AI technology develops, copyright law will likely need to be updated to address the unique challenges AI-generated art poses.