Who Owns AI-Generated Content? The Intellectual Property Debate in the Age of ChatGPT

The rapid advancement of artificial intelligence (AI) has transformed multiple industries, particularly content creation. AI-generated works, such as text, music and visual art, are everywhere, raising critical legal questions concerning intellectual property (IP) rights. Who owns the content produced by AI? Can AI be considered an author or inventor? Existing legal regulations to accommodate the emergence of generative AI technologies like OpenAi’s ChatGPT, leading to debates on authorship, copyright and the implications for human creators and businesses alike.

Copyright and AI-Generated Works

Copyright law traditionally protects original works of authorship created by human authors. Under the Berne Convention for the Protection of Literary and Artistic Works [1886], copyright protection is granted automatically upon the creation of an original work, provided that it reflects a human author’s intellectual effort. However, AI-generated content disrupts this paradigm, as AI lacks legal personhood and creative intention.

The UK Copyright, Designs and Patents Act 1988 (CDPA) is one of the few legal instruments that address computer-generated works explicitly. Section 9(3) of the CDPA states that “the author of a work generated by a computer shall be taken to be the person whom the arrangements necessary for the creation of the work are undertaken”. This provision implies that the individual who initiates or configures an AI system may hold the copyright, but its applicability to generative AI models such as ChatGPT remains contested. Unlike conventional software, where human involvement is more apparent, generative AI can produce content autonomously based on vast datasets, complicating the issue of human authorship.

International Perspectives and Emerging Jurisprudence

Jurisdictions outside the UK approach AI-generated works differently. In the United States, the Copyright Office has maintained a firm stance that copyright protection extends only to works with human authorship. In Thaler v Perlmutter [2023], the U.S District Court upheld the Copyright Office’s decision to reject AI-generated artwork’s copyright registration, reaffirming that AI-generated content cannot qualify for protection unless a human author exercises sufficient creative control.

In contrast, China has taken a more progressive approach, granting copyright to AI-generated content if human input is demonstrably substantial. In the landmark Shenzhen Tencent v Shanghai Yingxun [2019], a court ruled that an AI-generated news article was protected under copyright law, as it resulted from significant human guidance and editing.

The European Union (EU), through its Artificial Intelligence Act, has begun to explore regulatory frameworks for AI-generated content. While the EU Copyright Directive does not currently extend copyright to AI, discussions continue regarding potential reforms that balance innovation with the protection of human creators’ rights.

Ownership and Role of AI Developers

Another dimension of the debate revolves around AI developers and the companies that own AI models. Since AI-generated content derives from pre-existing datasets and training algorithms, technology firms such as OpenAI, Google and Meta could claim proprietary interests over AI-generated outputs. Terms of service agreements for AI platforms often dictate that users do not hold copyright over generated content. OpenAi’s terms, for instance, grant user permission to use ChatGPT-generated content freely but do not explicitly confer ownership rights.

This approach raises concerns about whether AI developers could indirectly control creative industries, limiting independent creators’ ability to monetise AI-generated works. Additionally, questions arise regarding potential liability - if AI-generated content infringes upon existing copyrights or produces defamatory material, who bears legal responsibility? These issues necessitate a nuanced legal response that clarifies ownership while ensuring accountability.

Patent Law and AI-Generated Inventions

Beyond copyright, AI-generated inventions present further challenges for patent law. The case of Thaler v Comptroller-General of Patents, Designs and TradeMarks [2021] in the UK, and its equivalent cases in the US and Australia, addressed whether an AI system named DABUS could be recognised as an inventor for a patent application. Courts in multiple jurisdictions, including the UK Supreme Court in 2023, have ruled that AI cannot be named as an inventor since patent law requires human agency. However, these decisions leave unresolved the question of whether AI-assisted inventions, where AI plays a substantial role in the creative process, should receive different treatment.

Ethical and Economic Implications

The uncertainty surrounding AI-generated content ownership has profound ethical and economic consequences. If AI-generated works remain unprotected by copyright, they risk becoming public domain material, disincentivising investment in AI-driven creativity. Conversely, granting excessive rights to AI developers could concentrate power in a few technology corporations, reducing fair competition.

Additionally, AI’s role in content creation challenges traditional notions of creativity and labour. Artists, writers and musicians increasingly face competition from AI tools capable of mimicking human expression. While AI-generated content can enhance productivity, it also raises concerns about the devaluation of human expertise and job displacement in creative industries.

Future Directions and Legal Reforms

To address these challenges, lawmakers must develop balanced legal frameworks that:

1. Clarify Human Authorship Requirements: Establishing clear guidelines on the threshold of human input required for copyright protection.

2. Define AI-Generated Works’ Ownership: Determining whether users, AI developers, or a combination of both should own AI-generated content.

3. Implement AI Liability Frameworks: Assigning responsibility for AI-generated outputs to prevent copyright infringement and misinformation.

4. Encourage Ethical AI Use: Promoting transparency in AI training data and ensuring AI systems do not exploit copyrighted material without permission.

As AI technology evolves, legal systems worldwide must adapt to maintain a fair balance between innovation, intellectual property rights, and the interests of human creators. While current law provide some guidance, the unique challenges posted by AI-generated content necessitate comprehensive reforms to ensure clarity, fairness and accountability in the digital age.

References and Further Reading 

"Berne Convention for the Protection of Literary and Artistic Works" (WIPO)

"Copyright, Designs and Patents Act 1988" (UK Parliament)

"Thaler v. Perlmutter" (U.S. District Court)

"Shenzhen Tencent v. Shanghai Yingxun" (China)

"Thaler v. Comptroller-General of Patents, Designs and TradeMarks" (UK)

"European Union Artificial Intelligence Act (Draft)" (European Union)

"Artificial Intelligence and Intellectual Property: Copyright and Patents" (UK Intellectual Property Office)

"Authors and Machines" (Columbia Journal of Law & the Arts)

"AI and Copyright: Reconciling the Human Authorship Requirement with Machine Creativity" (Harvard Journal of Law & Technology)

"Can AI Create Copyright?" (European Intellectual Property Review)

"OpenAI Terms of Service" (OpenAI)

"UK Government Report on AI and Intellectual Property" (UK Government)