When ChatGPT launched in late 2022, the conversations that followed were mostly about education, productivity, and job displacement. But quietly, a much older and more consequential conversation was being forced open: who owns what AI creates?
The question sounds simple. The answer is anything but.
For centuries, copyright law has rested on a foundational assumption: creative works come from humans. A person writes a novel, composes a song, paints a canvas, and the law grants them ownership because they made intentional creative choices. That framework worked beautifully for a world where humans were the only ones producing original work.
Generative AI broke that assumption. Tools like ChatGPT, DALL-E, Midjourney, and Suno can now produce texts, images, and music that rival human output in novelty and aesthetic quality. And they do it without intention, consciousness, or any of the attributes copyright law was built to protect.
The Academic Publishing Parallel
The disruption hit academic publishing early. When researchers started listing ChatGPT as a co-author on journal submissions, editorial boards scrambled to respond. Some journals now require authors to disclose any AI use in their writing process. Others ban AI-generated content outright. Many still haven’t published a clear policy at all.
The split reflects a deeper confusion about what authorship means when a machine contributes to the creative process. If a researcher uses AI to draft sections of a paper, then revises, reorganizes, and adds original analysis, is that fundamentally different from using a research assistant? And if so, where exactly is the line?
Academic publishing is still working through these questions. But at least the stakes there are relatively contained. In the legal arena, the same questions carry financial consequences worth billions.
A Global Patchwork of Legal Confusion
Bharati (2025) offers a thorough comparative analysis of how different jurisdictions are handling AI-generated works, and the picture is messy.
The United States strictly enforces human authorship. The U.S. Copyright Office has repeatedly denied registration for works created primarily by AI, holding that copyright requires a human author who made creative choices. If an AI generated the work and a human simply pressed “enter,” there’s no copyright protection.
The United Kingdom takes a different approach. Under UK law, authorship of computer-generated works is assigned to “the person by whom the arrangements necessary for the creation of the work are undertaken.” That’s vague enough to cover a range of scenarios, from the developer who built the AI to the user who crafted the prompt.
The European Union has focused on transparency and disclosure, requiring that AI-generated content be labeled as such, but hasn’t fully resolved the ownership question. Meanwhile, jurisdictions in Asia, including India and Japan, are handling cases on an individual basis, with inconsistent outcomes depending on the domain and the level of human involvement.
As Bharati argues, this fragmentation creates real problems for creators and industries operating across borders. A work that qualifies for copyright protection in London might have no legal standing in Washington.
Why Existing Attribution Models Fall Short
Part of the challenge is figuring out who deserves credit. Bharati identifies several attribution models that have been proposed, and explains why each one falls short on its own.
A developer-centric model gives rights to the people who built the AI system. A user-centric model gives rights to the person who prompted it. An investor-centric model assigns ownership to whoever funded the development. Each captures one piece of the creative process, but none reflects how creativity actually flows through a generative AI pipeline. The system design, the training data curation, the prompting, the selection of outputs, the post-generation editing: creativity is distributed across all of these steps, and no single actor controls the full chain.
According to Bharati, this is why courts are increasingly moving toward hybrid authorship models that recognize human-AI collaboration. Since 2022, judicial decisions have shown a clear trend toward acknowledging shared creative processes, even though AI itself has no legal personhood. The question is no longer “did a human or a machine make this?” but “how much meaningful human creative input went into the final work?”
A Proposed Framework
To address the authorship vacuum, Bharati proposes what he calls the Contributory Value Framework. The idea is straightforward: measure human creative contribution across five dimensions, including algorithm design, training data curation, creative direction, output selection, and post-generation modification. Rights would then be allocated proportionally based on how much meaningful creative input the human contributed at each stage.
For works where human involvement is substantial, standard copyright protections would apply. For works where AI operated with minimal human guidance, Bharati recommends a sui generis category: a separate legal classification with shorter protection terms, limited rights, mandatory registration, and required disclosure of AI involvement. This avoids the awkward exercise of forcing fully autonomous AI outputs into a copyright system that was designed for human creators.
Why This Connects to Education
If you’ve been following my recent posts, you might notice a thread connecting this to classroom concerns. The cognitive surrender research by Shaw and Nave (2026) showed that people accept AI outputs with minimal scrutiny, bypassing their own reasoning. That phenomenon plays out in creative and intellectual work too. When a student, a researcher, or a content creator accepts an AI-generated output wholesale, the question of who “authored” that work becomes genuinely difficult to answer. If no meaningful human judgment shaped the final product, the legal framework Bharati describes would say: no one did.
That has implications for how we teach students to interact with AI. If we want them to retain genuine authorship over their work, we need to teach them to engage with AI outputs critically, to select, revise, challenge, and reshape what AI gives them. The creative contribution has to be real, and it has to be demonstrable.
The Urgency of Reform
Bharati’s central argument is that policy reform can’t wait. Without legislative clarification, judicial guidance, and international coordination through bodies like WIPO, the uncertainty will continue to delay commercialization, distort incentives, and undermine trust across creative and legal industries.
The technology has moved. The law hasn’t. And every month that gap widens, the harder it becomes to close.
Reference
Bharati, R. K. (2025). Intellectual property rights in AI-generated creative works: Human authorship in automated production. Indian Journal of Law and Human Behavior, 11(2), 77–92. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6120486