DL Seminar | Protecting Amateur Creativity in the Age of Generative AI

Individual reflections by Priyana Aragula and Aris Huang (scroll below).

By Priyana Aragula

Cornell Tech

In a world that has been rooted in what we now call “traditional” methods of content generation, generative AI is making waves as an accessible alternative to these methods by users across the board. 

In her talk on 14th March, 2024, Kate Geddes (Postdoctoral Fellow at the NYU School of Law and the Digital Life Initiative at Cornell Tech) began with her motivation for the study of protecting amateur artists who produce content through the use of generative AI. She referred to her talk as a “A Love Letter to Remix Culture”, which indeed it was. 

“Remix Culture” was best explained in the talk through a comical video of an AI generated remix of the Lord of the Rings. Fans of the franchise used generative AI to create a meme by imputing Arnold Schwarzenegger’s face and voice onto characters in a scene. She went on to remark that generative AI has the capacity to promote remix culture by allowing users to engage interactively with cultural works and repurposing them as an expression of creativity.

It is not news that generative AI has had intellectual property issues in the recent past as a result of large corpuses of copyright protected work being used as training data for these models. Kate Geddes reasoned that amidst these growing concerns, the end users are neglected and receive relatively little attention. Some of her reasons included users generating the smallest economic footprint, general disdain for amateur creativity using AI, copyright being a law of author’s rights and not users’ rights, and more.  An interesting case made was how users are seen as “thieves”  and “pirates” of content, whereas authors are romantically viewed as “creative geniuses” by disregarding their capitalistic motivations. She went on to say that it is because of this moralistic language, that courts are careful to avoid undermining the rights of authors.

The importance of users rights was supported by examples of amateur artists such as Alex Smith and Rochelle Brock who use AI like Mid Journey to generate content in an attempt to address and overcome plus sized, queer, and people of color representation in mainstream cultural texts. The articulation of identity and community by modifying culture reworks was argued to be important for self expression and the development of a sense of self worth, competence and belonging. She also touched on semiotic democracy, explaining that the capacity to remix works promotes democratic value since culture texts are the terrain on which political values are formed. The conclusion being that it is important for everyone to have the ability to contribute to culture texts.

The second half of her talk introduced a variety of statutory solutions, along with pros and cons of each, that could be explored to protect user rights and expression. A few key points made during this discussion outlined the non-commercial use for the provision of generative AI, fair use being unreliable, existence of discrete carve outs for AI generated content for “less-valued” social causes, and compulsory licensing for commercial uses that allow artists to cover an existing song without liability. She also went over using a Digital Millennium Copyright Act (DMCA) style sage harbor. The DCMA allows platforms like YouTube to host content without being sued contingent on them taking down copyrighted work as soon as they are made aware of the same. She argued that a similar scheme could be used for generative AI. However, she made it clear through a plethora of cons such as content ID matching, that this model may not work seamlessly with generative AI.

She concluded her talk by acknowledging the difficulty and intractability of the problems surrounding protecting amateur creativity in the age of generative AI. On one hand, users who are otherwise incapable of producing content through traditional methods, can rely on generative AI for the same. On the other hand, there is a lot of concern around traditional artists such as authors, photographers, painters, etc, being replaced. Her ending statement reiterated that while there are various interests in this space that need to be reconciled, her study focuses heavily on the end users of generative AI. 

As informed by her talk, there are many aspects of users’ rights protection that raise questions of liability, copyright infringement, etc that do need to be addressed in the generative AI space. How do we measure the tradeoffs in each statutory solution discussed? Should courts be allowed the final say setting industrial policies? Who is to say the difference in cataloging a disabled individual using generative AI for self expression versus an individual with the lack of artistic ability using generative AI for capitalistic intent, as an “amateur artist”? While there isn’t one “correct” solution to this problem, it was evident that society continues to try and find a balance that allows us to accept generative AI in all its use cases today.

By Aris Huang

Cornell Tech

In the March 14 seminar, "Protecting Amateur Creativity in the Age of Generative AI," Kate Geddes presented a comprehensive overview of the challenges and opportunities presented by generative AI technologies, focusing on the impact on amateur creators. Geddes highlighted the promotion of a remix culture as a positive outcome of generative AI, using the imaginative example of "The Rings" featuring Arnold Schwarzenegger, created through deep learning, to illustrate her point. This example underpins the seminar's central thesis: generative AI holds the potential to democratize creativity, allowing individuals to remix and reimagine content in ways previously unimaginable.

However, Geddes critically analyzed the legal and ethical landscape surrounding amateur creativity in the digital age, noting a significant gap in the protection and promotion of user interests. She pointed out the systemic neglect of amateur creators due to their smaller economic footprint, the disdain for amateur creativity, and the predominant view of copyright law as a protector of authors' rights rather than users'. This has led to a restrictive environment where users are often labeled as "thieves" and "pirates," stifling creative exploration and sharing.

The seminar delved into the implications of overprotecting copyrighted works, which results in collateral censorship and discourages the widespread engagement of individuals in creative activities. Geddes argued that such an environment is less democratic and limits the potential for a vibrant semiotic democracy, where remixing cultural works helps in articulating identity and community. She highlighted the importance of user rights, not only for individual expression but also for contributing to the cultural dialogue that shapes public discourse.

To address these challenges, Geddes proposed several statutory solutions, including a non-commercial use provision and a compulsory licensing scheme, aimed at protecting user expression while ensuring that copyright owners receive some compensation. She stressed the need for a balanced approach that recognizes the value of user-generated content and supports the rights of amateur creators in the age of generative AI.

The content of Geddes' talk is particularly relevant in the context of the larger digital life landscape, where the lines between creators and consumers are increasingly blurred. Generative AI technologies have the potential to further democratize the creation of digital content, allowing more people to express themselves creatively and participate in cultural production. However, this potential is currently hampered by copyright laws and policies that favor corporate interests and established creators over amateurs and users.

Geddes' seminar serves as a call to action for a more inclusive approach to copyright and creativity in the digital age, urging for legal reforms that recognize and protect the rights of all creators, including amateurs. By fostering an environment where everyone can freely participate in the remix culture and contribute to the digital commons, we can ensure a richer, more diverse cultural landscape. This requires a collective effort to rethink copyright laws and policies to accommodate the realities of the digital age and the innovative possibilities offered by generative AI technologies.