(GA) AI Is Learning from Artists Without Permission: Who Truly Benefits from this blight?

GENRE 1: (GA) 

Last year when my choir performed Eric Whitacre’s Cloudburst, the room felt electrified. Every moment was stagnant as we realized all of our hard work and dedication finally paid off. The music is extremely technical as it requires harmonic understanding, breath control, and emotional connection. While the piece itself is inherently difficult, it also requires depth far beyond musical convention; in other words, it requires true connection to the music in order to be performed properly. The text is centered around Octavio Perez’s poem, The Broken Water Jar, a text entirely set in spanish, this represents the renewal of our planet, our minds, and our hearts. Coupled with the enriching harmonies and percussion, the performance was truly an otherworldly experience. No matter how advanced artificial intelligence becomes, it will never properly mimic the power that comes from the human voice. AI, in short, is just artistic mimicry without soul. It rhymes without reason and it is a rhythm without a pulse

Generative AI models are trained on enormous datasets, many of which scrape inspiration from copyrighted performances, vocal recordings, scores, writings and creative projects without the permission of these artists. In other words, AI companies are learning from artists, in the same sense that a student studies under their teacher. While this is a worthwhile comparison, there is one striking distinguisher: The “teachers” in this case are corporations worth billions of dollars, and the “students” never agreed to their works being used. 

Artists are already seeing consequences. AI generated songs using cloned celebrity voices are going viral online, without permission or compensation to artists. Visual artists find tAI-generated art mimicking their artistic style online. Writers have discovered that AI has completely replicated their voice.

Proponents of AI claim that training on copyrighted material falls under “fair use,” a legal defense meant to protect education and commentary. But scraping together the entire internet to fuel commercial models, and potentially gain millions of dollars in revenue, is not commentary. It is theft. More importantly, it erodes the dignity of artistic labor. When we listen to a choir perform Cloudburst, we feel something because humans are behind it, not a machine. AI models can replicate vowels, tone, or pitch, but it cannot replicate the shaping of sound, technical breathing, or human artistry. 

We need updated copyright laws that require AI companies to obtain permission before training their models on creative works. We need transparency about this data, and we need the public to reject  AI artists that rely on stolen human creativity. In an era of technological innovation, it’s important that we cherish what truly makes humans human. Art is one of the few things that makes us undeniably human, fueled by our own experiences, heartbreak, grief, and joy. If we allow AI to take it without asking, we risk losing more than jobs. We potentially risk losing creation itself. 

Video Performance of Cloudburst: https://www.youtube.com/watch?v=67O--CvmcpI

Genre 2: PSA POSTERS (Artists & Performers)

Genre 3: Policy Memo 

To: Members of the Committee on Intellectual Property and Emerging technology

From: Keith Williams, Vocal Performance, and History Major; researcher AI and Creative IP

Date: 11/22/25

Subject: Strengthening Regulations and protecting artists 

Executive Summary:

Current U.S. copyright does not adequately address copyrighted artistic materials in AI training datasets. As a result, AI companies can legally ingest, store, and profit from massive quantities of creative works; musical recordings, performances, vocal samples, scores, visual art, and literature, without creators’ consent. This regulatory gap has led to unauthorized voice cloning, artistic replication, and loss of income from artists. This memo outlines the potential risks of unregulated AI training practices and proposes policy framework that ensures consent, transparency, and compensation for creators. 

Background

Generative AI systems rely on enormous sets of datasets that frequently include copyrighted works scraped from the internet or purchased through data brokers. These works are used as a way to “train” models. Ultimately, this allows AI platforms to analyze patterns, styles, and artistic techniques. 

Because U.S. Law does not explicitly classify dataset integration as a form of copying. AI companies often argue that their training processes are protected under fair use. However, unlike traditional fair use, which involves critique, education, or transformation, AI training is a commercial activity. Training is performed on a scale that uses human-created works to generate computing outlets. 

Key Problems Identified 

• Lack of Consent: Creators are not asked before their works are used to train datasets

• No Compensation System: Artists receive no form of payment, even when their styles are being replicated for commercial use

• Ambiguous fair use Doctrine: Current fair use law was not designed for large-scale, commercial machine learning. 

• Threat to Creative Labor: Unrestricted AI imitation devalues human creativity and destabilizes artistic industries. 

Recommendations 

1. Mandatory Opt In Licensing System

Create a federal requirement that AI developers obtain licenses for any copyrighted content included in training datasets. Artists should be able to opt out by default and give consent explicitly. 

2. Data Transparency Regulations

Require AI companies to provide detailed public documentation of training datasets, including: 

• Sources of Data

• Copyright Status 

• Methods of Acquisition

• Whether materials were licensed or scraped

• Compensation and Royalty Framework

• Develop a royalty system, similar to streaming royalties, that ensures artists receive payment proportional to how their work contributes to AI model performance and output. 

• Federal Oversight of AI training Data

• Establish a federal body (Or extend Copyright office) to:

•  Audit AI training datasets

•  Enforce Consent and licensing standards

• Investigate violations

• Penalize noncompliant companies 

Conclusion: 

Generative AI has the potential to coexist with human creativity, but only if federal policy protects the rights and labor of artists. Implementing clear regulations around consent, transparency, and compensation will uphold the dignity of creative works. 

Rationale: 

This multi-genre project explores the ethical, creative, and legal consequences of Generative AI systems trained on the works of artists, musicians, and writers. By presenting the issue through three distinct genres: an op-ed, PSA poster, and policy memo, the project communicates through different audiences and uses varied rhetorical strategies to build an argument, AI may be technologically advanced, but its heavily unregulated use of human creativity harms artists, dilutes the value of artistic labor, and demands legal reform. 

Genre 1: (Opinion Essay) 

Purpose: 

The op-ed uses personal narrative and emotional appeal to introduce the central issue: that generative AI models lack the emotional depth and human experience that truly define art. By recounting the performance of Cloudburst and highlighting the spiritual, technical, and emotional dimensions of choral singing, the essay establishes what is at stake when AI replicates artistic works. 

Audience: 

The general public, readers of newspapers, magazines, or online opinion columns who may not fully understand how AI training is impacting artists 

Rhetorical Choices: 

Pathos: Personal storytelling to convey emotional investment and the irreplaceable quality of human art.

Imagery: Descriptions of Harmony, breath, and the atmosphere of performance help readers feel the difference between human expression and AI mimicry.

Contrast: AI is portrayed not as creative, but as mimicry without soul

Ethical Framing: This essay positions unauthorized training not as a legal problem, but as a moral one that violates artistic dignity. 

Genre 2: Public PSA poster

Purpose:

This poster functions as a public awareness tool designed to quickly educate and motivate viewers. Unlike the opinion essay, which requires sustained reading and listening, the PSA uses vibrant colors to attract viewers, and short statements to educate viewers as well. 

Audience:

Artists, performers, writers, teachers and general consumers on social media or in physical spaces. 

Rhetorical Choices: 

Concise, Impactful Imagery: Slogans like “AI is taking away from artists, not helping them” 

Visual Urgency: Imagery that symbolizes demonization, and theft

Ethos & Pathos: Appeals to artists’ pride in their craft and desire for autonomy 

Call to Action: Encourages viewers to stay informed, vigilant, and mindful

Genre 3: Policy Memo 

Purpose: This policy memo addresses the structural, legal side of the issue and proposes concrete solutions: Licensing systems, transparency requirements, updated fair-use doctrines, and federal oversight, This shifts the conversation from emotional appeals and public awareness to legislative action. 

Audience: Members of the Committee on Intellectual Property and emerging technology, lawmakers, and policymakers

Rhetorical Choices:Logistical Organization: Executive summary, background, problems, recommendations, and conclusion

Professional, objective tone: Establishes credibility as a researcher and musician

Evidence- Based Claim: Describes specific harms artists experience such as voice, cloning, unauthorized imitation, and loss of income.

Sources: 

“Colonizing Art.” OpenMind Magazine, 2022, www.openmindmag.org/articles/colonizing-art?gad_source=1&gad_campaignid=20511855727&gbraid=0AAAAAocyGTYDJhXf7szgMf1ScLeoBcw5z&gclid=Cj0KCQiAxJXJBhD_ARIsAH_JGjhjPKN51CUZEumoLWCn_1EQ4o_bBMzP03XUuVGihCR90b2855j5B3caAgoaEALw_wcB. Accessed 25 Nov. 2025.

Wagner, Jennifer. “To Ingrain AI Ethics, We Should Get Creative about Copyrights.” Undark Magazine, 13 Apr. 2023, undark.org/2023/04/13/to-ingrain-ai-ethics-we-should-get-creative-about-copyrights/?gad_source=1&gad_campaignid=18401135165&gbraid=0AAAAADKKeAw7Spv2Sv-l5EuB1vMOejWR4&gclid=Cj0KCQiAxJXJBhD_ARIsAH_JGji6nvIG5XeQ3X12Bp17GwT-jadhUnLcvFtbgm3IplFQ7n8P0JHJ5t4aAo-2EALw_wcB. Accessed 25 Nov. 2025.

‌ “Artificial Intelligence Consulting and Advisory.” EisnerAmper, 2025, www.eisneramper.com/services/advisory/artificial-intelligence-consulting/?utm_term=ai%20compliance&utm_campaign=Artificial+Intelligence+Services+-+Search&utm_source=adwords&utm_medium=ppc&hsa_acc=1797353098&hsa_cam=22532358369&. Accessed 25 Nov. 2025.

France, Lisa Respers. “Xania Monet Is the First AI-Powered Artist to Debut on a Billboard Airplay Chart, but She Likely Won’t Be the Last.” CNN, Nov. 2025, www.cnn.com/2025/11/01/entertainment/xania-monet-billboard-ai.

‌