Artificial Intelligence and Intellectual Property Rights

The scope of AI has expanded beyond autonomous vehicles, healthcare, and education; it now autonomously produces artistic creations and devises several new products and services for which various intellectual property rights can be claimed if developed by humans. Currently, no legislation grants Intellectual Property Rights for inventions or artistic creations generated by AI. They are not products of human mind but rather creations of “AI intellect.” The eligibility of AI-generated works for Intellectual Property Rights remains a subject of controversy, as does the determination of ownership in such cases. Initially, individuals believed that the alterations introduced by AI would be minimal and that society could assimilate it without any adjustments to the existing intellectual property rights framework. Nonetheless, the actuality diverges. Technological breakthroughs in AI have transformed the notion of evolution, leading to concerns that AI may surpass human capabilities. The advancement of AI technology is akin to science fiction. In this context, it is imperative to revise the current intellectual property rights legislation.

Typically, the author or innovator is granted initial ownership rights, and most jurisdictions mandate that the author or inventor be a natural person. These rights are assignable by authors and inventors. When employees create an invention during their job, the invention is owned by the corporation, as it falls under the ‘work for hire’ doctrine or the employment contract. In fact, the majority of patents are held by artificial entities, specifically corporations, rather than individual inventors. Even when enterprises hold intellectual property rights, to acknowledge the rights of the human inventor, the legislation of most nations mandates that a natural person be identified as an author or inventor. These Intellectual Property Laws were not created with the advancement of AI in mind. In the majority of AI-generated creations, there is an absence of a competent human inventor or author. The employer-employee relationship bears significant resemblance to the relationship between AI and its owner or developer.

CONTEXT

At present, the quantity of AI systems independently inventing is rather low; in most instances, AI serves merely as a tool. If AI is utilized solely as a tool, it cannot be regarded as the inventor. A command is issued to a computer program created by a human to resolve a complex algorithm; if successful, the program remains a tool rather than the inventor. Conversely, if an AI-driven machine is tasked with autonomously developing an improved version of a product, it may be regarded as an inventor. A policy challenge emerges about the patentability of AI-generated inventions in the absence of a human inventor. The primary question in such instances is who should be granted ownership of the patent. This situation remains ambiguous as of 2023. This condition is unwelcome as it allows individuals to claim credit for AI-generated work by engaging with patent offices. AI will not contest the attribution of credit to humans for its outputs; nonetheless, this situation is profoundly unjust to individuals who are authentically innovating without the assistance of AI.

Regarding copyright, more laws can be identified. The United Kingdom was the inaugural nation to safeguard computer-generated works as early as 1988. In the UK, the “producer” of the work is regarded as the author and is bestowed with copyright privileges. The Copyright Office in the USA rejected AI-generated music in 1950. In 2019, the US Copyright Office rejected a proposal to grant an AI the ability to copyright a work of art. A three-member panel evaluated a 2019 ruling concerning Stephen Thaler, who sought to copyright a photograph on behalf of his Creativity Machine algorithm. The tribunal concluded that Thaler’s AI-generated image lacked “human authorship,” which it deemed a requisite criterion for protection.

The aforementioned work by Creativity Machine is entitled “A Recent Entrance to Paradise.” Thaler, who sought copyright for it, describes it as a “simulated near-death experience,” wherein an algorithm reconstructs images to produce hallucinatory visuals and a fictional narrative regarding the post-mortem experience. The AI accomplishes this with minimal human intervention. The board stated in the ruling that a fundamental requirement for copyright registration is the existence of a connection between the human intellect and creative expression, therefore precluding nonhumans from receiving copyright protection. In the USA, AI-generated works are classified as public domain, allowing anybody to exploit them and claim credit for the AI’s output.

In the historical case of 1884, Burrow Giles v. Sarony, the photographer Sarony[1] initiated a lawsuit against the Burrow Giles Lithographic Company for copyright infringement concerning a renowned portrait of Oscar Wilde. The company asserted that the photographer could not be deemed the author of the photograph, as it is merely a mechanical reproduction of a natural phenomenon captured by a camera; however, the US Supreme Court countered this argument, ruling that any medium “through which the ideas in the author’s mind are rendered visible” is entitled to copyright protection. Consequently, the case addressed two principal issues: may human authorship be negated due to the utilization of a machine, and may a camera or a machine be regarded as an author? Prior to 150 years ago, no one envisioned that a machine could possess the capacity for thought. Copyright laws necessitate original concepts and intellectual creations. It was posited that machines lack these attributes; hence, they cannot be afforded copyright protection. The current landscape has markedly transformed, as deep learning and machine learning enable artificial intelligence to produce creative concepts akin to human cognition.

[1] Burrow-Giles Lithographic Company v. Sarony, 111 U.S. 53 (1884)