Part 2: Should You Be Able to Copyright Works Made by AI?

Poudre Press Staff
4 hours ago
9 min read

The growing use of AI and its increasing acceptance also raise risks and ethical questions. Due to the complexity of this issue, it will be divided into two parts: Part One, written by Naomy Vazquez-Alvarez, arguing that AI works should be copyrightable; and Part Two, written by Simon Belda, arguing that AI works should not be copyable. The stance we take in each article may or may not be the stance we actually believe in.

In the current political climate, AI companies are spending a lot of resources on AI, which includes over "$252.3 billion in 2024 alone," says Stanford HAI, without recouping any of this cost. To keep this number low, many companies have refused to cooperate with websites and articles that do not want to be read by AI, as well as copyright holders who are often overlooked when the AI reproduces an article. Despite the limited use case for generative AI, granting it copyright protection would be unfair to artists and would also remove all meaning from existing copyrights. Yet, as AI systems become more powerful and pervasive, serious ethical and legal questions arise. One of the most consequential is whether works generated by AI, such as text, code, images, or music, should be eligible for copyright protection. This essay will argue that they should not be. Granting AI-generated works copyright would undermine the fundamental purpose of copyright law, disadvantage human creators, and destabilize the balance designed to encourage creativity while protecting original authors. Despite the purpose of copyright law and why AI-generated works do not fit, the unfair competitive advantage that AI-generated copyrightable works would give to wealthy tech entities, the moral and ethical dimension of creativity and ownership, and possible alternative frameworks to address AI-generated content without granting full copyright.

Some say AI-generated works are creative and new, so they deserve protection. AI outputs can indeed be novel because no identical text, image, or music has existed before. But novelty alone is insufficient. Copyright jurisprudence typically requires more than novelty; it requires creative choices, expression, and original authorship. AI lacks the subjective intentionality and creative judgment that define human art. Granting copyright solely based on novelty divorces the law from its foundational principle: protecting human creative labor. Furthermore, acknowledging AI outputs as copyrightable would incentivize quantity over quality, scale over substance.

Did you know the first AI chatbot, ELIZA, was created from 1964 and 1967, nearly 50 years before popular assistants like Siri and Alexa existed

Corporations could flood the market with AI-generated content, undermining the value of human-created works and reducing art to an industrial output. Without copyright, there is no incentive to create high-quality AI outputs. This argument treats AI outputs as akin to human-created works, but that conflates fundamentally different things. AI outputs are the result of computational processes; they do not require human effort in the moment of creation, beyond writing a prompt or pressing a button. The “investment” has already been made at model-creation time; the marginal cost of generating each new piece is often negligible. To grant copyright would effectively reward a zero or low-effort process, displacing the incentive structure that copyright intends to support. Moreover, even if some form of incentive is needed for AI developers, that incentive can be provided through other legal regimes, patents for novel model architectures, trade secrets, licensing agreements, without conflating those with creative authorship. Some also bring up that AI-generated content can benefit society — more books, art, music. There is merit to the idea that AI could democratize content creation, lowering barriers and enabling wider access. But unregulated proliferation risks overwhelming human creative ecosystems. Moreover, content produced by AI may lack depth, authenticity, and the human connection that gives art its true value. If society values creative work as more than a commodity, as a reflection of human experience, then we must resist treating AI-generated content as equivalent. Instead, we should encourage AI as a tool to augment human creativity, not to replace it wholesale.

Copyright law historically has two key purposes. First, it incentivizes creative effort by granting authors a limited-time monopoly over their works. This allows creators to profit from their labor and encourages ongoing innovation in the arts,

Did you know, the AI used in writing models or LLMs is predicting what words will come next, rather than actually thinking

literature, science, and other domains. Second, it protects the moral and reputational interests of authors — their right to attribution, to the integrity of their work, and to control how it is used. At its core, copyright rewards human labor: the expression of ideas through skill, judgment, originality, and creativity. Human creators bring intention, experiences, cultural context, aesthetic sensibilities — all shaped by the complexities of human life. By awarding copyright to human authors, the law recognizes and protects the uniquely human dimension of creative labor. By contrast, AI systems do not "experience" or "intend" in the human sense. They manipulate patterns learned across massive datasets; they are not conscious, nor do they have subjective experiences, motivations, or moral agency as of now. Their outputs are statistical byproducts, not expressions of personal lived experience. To grant such outputs the status of copyrighted works would stretch the meaning of "authorship" beyond recognition: from creator to algorithm. The traditional test for copyrightability demands originality, that is, creative choices not wholly dictated by another source. Even when a human remixes or transforms existing material, as long as there is meaningful human input, the resulting work may be protected. But AI-generated works, as they are produced by automated processes based on training data, lack genuine human creative judgment. The "originality" is a derivative of the enormous compilation used to train the model. If we allow AI-generated works to be copyrighted, we risk collapsing the boundary between original creation and derivative regurgitation. The piece becomes a byproduct of pattern replication rather than authentic creative expression. Copyright also acknowledges the moral rights of creators, their connection to the work as human beings. Granting copyright to AI-generated works severs that moral link. It would put ownership in the hands of corporations or developers who deploy models, not human creators. This dilutes the human dimension of authorship, reducing art and literature to algorithms and code, which is a material representation of expression that undermines human creativity.

The massive surge in AI investment, and as said before, $252.3 billion dollar AI bubble in 2024 (Stanford HAI), shows the scale at which AI is being developed and deployed. That investment reflects not only research and development costs, but also resources poured into infrastructure, data centers, personnel, and marketing. For companies that invest at this scale, granting copyright to AI-generated works would offer tremendous financial upside. They could flood media, literature, art, music, technical documentation, and any number of creative domains with AI-generated copyrighted content, effectively capturing and removing entire markets. This creates a structural advantage for large corporations with the resources to invest(Negar Bondari, 2025). Human creators, independent artists, freelance writers, musicians, and educators generally do not have access to the massive computational resources, data, and infrastructure of large AI firms. If AI-generated works are copyrightable, corporations would be able to produce vast quantities of creative material at near-zero incremental cost and claim ownership. Meanwhile, human creators would struggle to compete, unable to match the speed or volume. This would distort the creative ecosystem, shifting it away from human-centered art toward algorithm-churned content optimized for scale and profitability. The result: a form of content monopolization by AI companies. If AI-generated works are copyrighted, there is also a risk that existing human-created works lose value. Consider this: when the market becomes saturated with algorithmically generated images, texts, and music, what becomes of the worth of a human-made novel, a painting, or a song? The uniqueness, rarity, and human touch that give these works value would erode. Moreover, authors whose works (or styles) were used to train AI models might find their own creative

outputs devalued, or (as seen below) altered in ways they did not want or remove the message the author was attempting to convey. Not to mention the legal ambiguity about derivative works, attribution, and whether their consent was obtained for training before the AI consumed it.

Existing copyright doctrine often requires human authorship, originality, and expression. Historically, courts and copyright offices have been reluctant to recognize non-human authors as rightful copyright holders. “In many jurisdictions, copyright protections have been applied only to works created by humans” (Andrés Guadamuz, 2017). This is because copyright law is built on the notion that creative labor is a human activity. Granting AI-generated works copyright would represent a radical departure from this tradition. Without a human author, it is unclear who would hold the copyright. The company that trained or ran the model? The model itself? The user who provided the prompt? The legislation has not evolved to handle such cases(Kate Knibbs 2024), and for good reason. Even if the law were changed to allow copyright for AI-generated works, attribution would be problematic. AI outputs are derivative from countless human-created works used during training; as Kate said, “what portion is novel, what is derivative, and who should be credited could be legally intractable”(Kate Knibbs 2024). ass seen in the image below, it can, will, and does directly copy copyrighted works.

If every prompt that yields a text, image, or music piece is automatically copyrighted, the volume of new copyrighted works would explode dramatically. The administrative and legal burden of tracking, licensing, and policing would be enormous. Finally, such a change would favor large corporations with legal and financial resources; small creators and independent artists would likely be marginalized.

Given the scale of AI investment and the power of generative systems, it is unrealistic to expect AI to disappear in these markets. Instead, we should consider alternative methods that strike a balance to protect human creators while allowing for responsible use of AI outputs. Rather than granting full copyright, governments could create a new, special category of rights for AI-generated works. Such a regime might treat AI outputs more like data or databases. granting limited, time-bound, and non-exclusive usage rights, or requiring licensing that ensures fair compensation to human creators whose work contributed to the training data. Such a system would acknowledge that AI-generated works are different in kind from human-created ones, and avoid conflating machine outputs with human creativity. AI developers and deployers should be required to maintain detailed logs of training data, prompt-output mapping, and usage. If an AI-generated work is used commercially, the models, the datasets, and (as far as possible) the human origin sources should be disclosed. This would help preserve accountability, guard against misuse, and help ensure that individuals whose work contributed to training are not exploited or ignored. Because AI models learn from the creative labor of countless human authors, musicians, artists, journalists, and others, there is a moral argument that those contributors should be compensated, especially if their work leads to commercial outputs. Licensing frameworks could be developed that allow contributors to opt in or out of training datasets, similar to the previous methods of removing bots from your website. Alternatively, collective-licensing schemes could be created to pool royalties for creators who consent. similar to how music licensing operates today, or how stock photo agencies manage rights. Even if AI-generated works are used, human creators whose original works were used for training should retain rights: the right to attribution, to prevent derogatory or misattributed use of their work, and to contest derivative works that greatly distort their contribution. This would help preserve human dignity and ensure that AI development does not amount to a free-for-all appropriation.

The decision about whether AI-generated works should be copyrighted is not a small technicality. It is a foundational moral, cultural, and economic choice. The stakes are high. Human creativity comes in many forms. These are not just products; they are part of our shared humanity. Granting copyright to AI works threatens to drown out real voices under a flood of machine-generated, algorithm-optimized content. A world where AI-generated works dominate could lead to cultural uniformity, where the subtlety, nuance, and diversity of human voices are lost. Unique perspectives. marginalized communities. underrepresented cultures. And individual artists risk being replaced by generic, mass-produced content optimized for virality, speed, or profitability. Copyright exists, in part, to ensure that creators, writers, musicians, and artists can make a living from their work. If AI-generated works are copyrighted, the economics shift dramatically in favor of large entities with access to data, computing power, and capital. Independent creators would likely be marginalized or economically squeezed out. This is not just about fairness to individual artists; it's about preserving a pluralistic cultural economy. If only a few powerful firms dominate creative output, society loses not only diversity but the economic independence of creators. When humans create, they stake their identity, experience, and reputation on their work. They deserve attribution, respect for how their work is used, and control over derivative works. Granting copyright to AI outputs severs these moral connections. It makes authorship anonymous, commodified, and detached from its human origin. Moreover, it undermines transparency. When anyone can feed a prompt to an AI and produce a “copyrightable” work, attribution becomes meaningless.

| ; |

Dont end our creativity for us

Read the next article here! https://www.poudrepress.com/

Bibliography

Bondari, Negar, and Hidetaka Okamoto. 2025. “AI, Copyright, and the Law: The

Ongoing Battle Over Intellectual Property Rights – IP & Technology Law Society.” USC. https://sites.usc.edu/iptls/2025/02/04/ai-copyright-and-the-law-the-ongoing-battle-over-intellectual-property-rights/.

https://www.wipo.int/en/web/wipo-magazine/articles/artificial-intelligence-and-copyright-40141.

https://www.wired.com/story/ai-copyright-case-tracker/.

creativity, says US Copyright Office.” AP News. https://apnews.com/article/ai-copyright-office-artificial-intelligence-363f1c537eb86b624bf5e81bed70d459.

U.S. Copyright Office. 2023. “Copyright and Artificial Intelligence | U.S.” U.S. Copyright

Office. https://www.copyright.gov/ai/.

Zirpoli, Christopher T. 2025. “Generative Artificial Intelligence and Copyright Law |