Copyright Ambiguity of AI Artwork

The creative powers of AI are evolving rapidly, and the definition of the word "artist" needs immediate clarification.

NFT images by ThePicassoProject, a Picasso-mimicking AI made by UC Santa Barbara students

The Advent of Robotic Artists

The realm of the arts is drastically changing. Recently, I have seen Adele's "Hello" sang by Plankton and Matt Maltese's "As the World Caves In" sang by a sheep from Minecraft while scrolling TikTok. With the arrival of free or low cost AI art generators like OpenArt, I have seen creators fulfill their imaginative curiosities by merging Spiderman and Batman while others try to revive deceased artists like Picasso and Rembrandt by having the AI mimic their style. The majority of AI artists rely on neural networks, which are complex computing systems comprised of interconnected nodes modeled after the human brain. These networks demand large amounts of real data to grasp user-provided parameters and autonomously make decisions in the creative process. Neural networks and AI art are not new phenomenon, however, but advancements in deep learning and greater investments into the industry have resulted in much more powerful AI artists recently. With the continuous development of new AI projects and organization of different data, AI can now compose music, manipulate images, write poems, and generate videos with any style or property a user desires.

Coming into a Lighter Place by artist and programmer Harold Cohen, 1998, created via AARON, an early series of computer programs that creates original artistic images

The Question For Artists

The majority of legal systems, like in Spain, assert that copyright protection is exclusively granted to works that are produced by human creators. Before, there was no question of intellectual property regarding digital art because the computer software was used more as a tool. However, with artificial intelligence claiming more and more autonomy in the creative process, it no longer functions as a tool or extension drawn and new laws must be made to identify the owner of copyright. There seems to be three copyright systems that can be enforced. First, the artificial intelligence itself becomes the copyright owner and the criteria for copyright ownership needs to be amended. Second, the user, programmer, or AI company is the copyright owner. This would be an expansion of the "work for hire" doctrine, in which ownership of a copyright is transferred to the "employer." Lastly, there is no copyright owner for AI artwork, and the work would automatically enter public domain.

Any way in which the legal system addresses emerging forms of machine-driven creativity may carry significant commercial ramifications. For instance, if AI works are released into the public domain, a company behind an AI making logos or advertisements for other brands wouldn't be able to sell their work. Advancements in machine learning require millions, if not billions, of dollars, so if AI work isn't protected by law the industry may lose significant investment capital. Conversely, utilizing artificial intelligence for labor-intensive tasks may remain justified due to potential savings in personnel expenses. Currently, the US, the EU, and Australia reserve copyright protection specifically for human-created works. However, Hong Kong (SAR), India, Ireland, New Zealand and the UK gives authorship to the programmer of the AI. The UK copyright law, section 9(3) of the Copyright, Designs and Patents Act (CDPA), states:

“In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.”

Personally, I believe copyright should be granted to the programmer or company of the artificial intelligence; it ensures the company's investment and return on the technology and propels the industry forward.

The Future of AI Artwork

The future of AI artwork and its intersection with copyright policy is filled with turbulence. Currently, there are some notable distinctions between AI generated artwork and those created by humans. For examples, one can tell if a portrait is created by a machine if the hands don't look quite right. However, AI technology is improving at an immense rate and we are nearing the day in which we cannot distinguish between human-generated and machine-generated content. At the point, an option may be to deny copyright to any work involving artificial intelligence to protect and reward slower human artists for their skill, labour, and effort. Another looming issue is the security of media and public safety. AI is already very strong in generating realistic videos and deepfakes. In the future, people will need to be much more skeptical about the news they receive, as criminals may use false imagery to scare, provoke, deceive, and divide entire countries or even the world. This highly relevant subject could potentially exert significant influence and enter the national spotlight very soon, particularly in the upcoming 2024 presidential election. I hope to create another post regarding the security and legality of deepfakes soon, so stay tuned!

Midjourney-created images of Donald Trump being arrested that went viral in March 2023

Sources:

https://www.independent.com/2023/01/19/uc-santa-barbara-doctoral-students-build-neural-network-to-paint-like-picasso/

https://www.wipo.int/wipo_magazine/en/2017/05/article_0003.html

https://en.wikipedia.org/wiki/Artificial_intelligence_art