By Jerameel Kevins Owuor Odhiambo
The proliferation of artificial intelligence has transformed the landscape of digital content creation, with deep fakes emerging as a particularly disruptive force at the intersection of technology and intellectual property rights. These hyper-realistic synthetic media, created using sophisticated deep learning algorithms and generative adversarial networks (GANs), have experienced exponential growth, with DeepMedia estimating approximately 8 million deep fake videos circulating online by 2025. This dramatic expansion raises profound questions about the ownership of digital personas, the legality of training AI on copyrighted materials, and the adequacy of existing IP frameworks to address these novel challenges.
Current intellectual property jurisprudence struggles to accommodate AI-generated content, as evidenced by landmark cases establishing the human-centric nature of copyright law. The U.S. Copyright Office’s rejection of protection for AI-generated art in the 2023 “Zarya of the Dawn” case, coupled with the DABUS patent decision where the Supreme Court declined to recognize AI as an inventor, demonstrates the legal system’s reluctance to extend IP rights to non-human creators. This creates a significant regulatory gap for deep fakes that appropriate real individuals’ likenesses without authorization, leaving victims with limited recourse when their digital identities are manipulated.
The unauthorized use of copyrighted materials to train AI models represents another contentious frontier in IP law. The ongoing litigation between The New York Times and OpenAI/Microsoft exemplifies this tension, with the media company alleging that large language models infringe copyrights by ingesting vast quantities of protected journalistic content without proper licensing agreements.
Similarly, the 2023 Anil Kapoor v. Simply Life India case established a precedent that unauthorized deep fakes violate personality and publicity rights, reinforcing the necessity of consent when utilizing a person’s likeness for commercial purposes. These cases highlight the global divergence in approaches to resolving AI’s implications for intellectual property.
Beyond commercial concerns, deep fakes pose serious threats to democratic processes and public trust. During India’s 2024 Lok Sabha elections, fabricated videos showing popular actors Ranveer Singh and Aamir Khan endorsing political parties went viral, illustrating the potential for synthetic media to manipulate voter opinion. In the United States, a sophisticated deep fake robocall imitating President Biden’s voice resulted in a $1 million fine for Lingo Telecom for violating telecommunications regulations. These incidents have catalyzed regulatory responses such as the European Union’s Artificial Intelligence Act, which imposes transparency requirements for AI developers but must carefully balance these protections against established privacy and expression rights.
Financial institutions have become prime targets for deep fake-enabled fraud, with significant economic implications. The 2024 case of British engineering firm Arup losing $25 million to scammers using AI to impersonate its Chief Financial Officer during a video conference represents a troubling trend. According to the Financial Crimes Enforcement Network (FinCEN), losses from deep fake-driven financial fraud in the United States are projected to reach $40 billion by 2027, prompting the agency to issue alert FIN-2024-Alert004 urging financial institutions to implement enhanced detection protocols for synthetic media. These developments underscore the critical need for robust IP protections working in concert with advanced technological countermeasures.
The regulatory landscape addressing deep fakes and intellectual property varies significantly across jurisdictions. While the EU has implemented the risk-based Artificial Intelligence Act mandating transparency in AI training data, the United States has primarily relied on existing copyright frameworks supplemented by proposed legislation like the NO FAKES Act, which would establish federal rights over one’s voice and likeness. Other countries have taken different approaches, with Singapore and the EU introducing specific exceptions for text and data mining in AI development, and the United Kingdom recognizing “computer-generated” works as a distinct IP category. A 2025 OECD report highlighted these disparities, noting the challenge of harmonizing intellectual property regimes with AI’s inherently borderless nature.
The ethical dimensions of deep fakes extend beyond legal frameworks, particularly regarding consent and exploitation. A 2024 BBC investigation documented the alarming prevalence of non-consensual deep fake pornography on platforms like CivitAI, with vulnerable populations disproportionately targeted. Academic research by Albahar and Malki (2019) has characterized such privacy violations as “major ethical concerns,” while the precedent established in Naruto v. Slater suggests that new legal categories may be necessary to address ownership and liability for AI-generated content. Some scholars have proposed concepts like “Digiworks Rights” to fill this gap, recognizing the unique characteristics of synthetic media that existing IP frameworks fail to capture adequately.
Looking forward, addressing the complex interplay between artificial intelligence, deep fakes, and intellectual property will require adaptive legal frameworks complemented by technological innovation. The World Intellectual Property Organization’s 2020 proposal to expand authorship concepts to include AI-human collaborations represents one potential path forward, while investments in detection technologies, as advocated by a 2024 Crime Science review, provide another crucial component. As AI capabilities continue to advance, stakeholders must collaborate on cohesive strategies that protect individual rights and the integrity of our information ecosystem while fostering beneficial innovation. The challenge lies in striking this delicate balance—ensuring that our legal and ethical frameworks evolve in tandem with technological capabilities to create an environment where AI can flourish without undermining fundamental intellectual property protections or eroding public trust in digital media.
The writer is a legal writer and researcher