AI and Copyright Laws: Is Artificial Intelligence Eligible for Copyright Protection?
Artificial Intelligence and Copyright Law: A Blurring Line
Artificial Intelligence (AI) has revolutionized the creative landscape, with generative AI leading the transformation. However, the traditional copyright framework, built upon human authorship, encounters challenges when confronted with AI-generated content.
Currently, under existing copyright laws in most countries, only human authors can hold copyright. AI itself is not recognized as an author. This was reaffirmed by the 2021 Thaler case in the U.S., where the court held AI cannot be an author, but humans who use AI as a tool in their creative process can claim copyright if there is sufficient human involvement.
Some countries, including the UK, Hong Kong, India, Ireland, and New Zealand, grant copyright to the person who undertakes the arrangements necessary for creating the work. In practice, this is often interpreted as the AI developer or operator rather than the AI system itself.
In the U.S., the Copyright Office excludes non-humans from authorship, leading to uncertainties especially for fully AI-generated works. Court decisions such as U.S. District Judge William Orrick's upheld copyright infringement claims against AI companies training on copyrighted works without consent.
Regarding the AI training process, two California courts recently ruled that copying copyrighted written works to train large language models (LLMs) can qualify as "fair use" given the transformative nature of the training. However, this does not give carte blanche, as plaintiffs can still argue market harm or piracy.
Legal experts debate multiple approaches to AI authorship, including treating AI itself as the creator, granting copyright to the AI’s owner, assigning copyright to the AI programmer or operator who developed or trained the model, or considering humans who curate or prepare the AI training data as authors.
Increasingly, legal scrutiny focuses on training data licensing and liability. Several lawsuits by authors against AI companies argue for compensation due to unauthorized use of copyrighted works in training. The U.S. Copyright Office’s May 2025 report concluded that AI developers using copyrighted works to generate content that competes with originals go beyond fair use, signaling a potential tightening of protections for creators.
As technology advances, courts and legislatures worldwide are expected to continue refining definitions of authorship, fair use, and liability in AI-generated content. The regulatory landscape remains in flux, with proposed changes focusing on clarifying authorship definitions, ownership rights, and liability relating to AI's role in creating content.
Attributing copyright ownership becomes a paradigm shift when AI generates works devoid of direct human influence. There is a call for legislative contemplation and potential amendments to the law to grapple effectively with the challenges posed by AI in intellectual property protection. AI companies may assign copyright rights to users, highlighting the role of user agreements in shaping the contours of ownership.
The AI training process, which involves the use of copyrighted works, raises questions about its alignment with copyright principles. Encouraging ongoing discussions within legal and tech communities is essential to staying attuned to the transformative convergence between artificial intelligence and intellectual property. Generative AI programs like DALL-E and ChatGPT showcase AI's ability to autonomously generate diverse creative content.
The question of whether AI training constitutes fair use under copyright law is scrutinized, considering the four statutory factors: purpose, nature, amount, and market impact. Contracts and terms of service are crucial in defining copyright ownership for AI-generated works, as they often define whether the AI entity or the user assumes authorship.
The ownership of AI-generated copyright is a complex issue that requires examination of the creation process and the definition of authorship. Determining the rightful owner and establishing the creative lineage of AI-generated content introduces complexities in IP protection frameworks. The court ruling in Thaler v. Perlmutter affirmed the denial of copyright registration for an AI-generated artwork, asserting the fundamental requirement of human authorship for copyright protection. The decision in Thaler v. Perlmutter highlighted the historical context, emphasizing that copyright law traditionally seeks to incentivize and protect works of human origin. The traditional prerequisites of human-centric creativity struggle to accommodate the machine's autonomous creative processes.
Establishing access and substantial similarity is crucial in distinguishing copyright infringement for AI outputs, with the training data gaining prominence in legal scrutiny. As the role of AI in content creation continues to evolve, so too will the debates surrounding copyright protection and the legal frameworks designed to govern it.
- In the debate on AI authorship, questions arise about whether AI itself should be considered a creator, or if copyright should be granted to the AI's owner, the AI programmer, or the operator who trained the model.
- The AI training process, which might involve copyrighted works, presents a challenge for alignment with traditional copyright principles, as the question of whether it qualifies as 'fair use' under copyright law is scrutinized.
- Courts and legislatures worldwide are increasingly focusing on the issue of training data licensing and liability, with several lawsuits filed by authors against AI companies for unauthorized use of copyrighted works in training.
- As technology continues to advance and AI generates works with less direct human influence, a paradigm shift is needed in copyright ownership, spurring calls for legislative contemplation and potential amendments to the law in the area of intellectual property protection.
