AI-Generated Works – Research Note
Many of the actions that artificial intelligence (AI) can perform today fall within the purview of certain legal rights, such as those outlined in copyright law. However, during the inception and development of these laws, as well as their subsequent application and case law formation, AI was not as actively involved. Consequently, the lack of recognition of artificial intelligence by traditional legal systems has led to a situation where works created by AI are not recognized in the US.
Envisioning the potential future advancements of rapidly evolving technology is challenging. Nevertheless, it is foreseeable that AI technology will play an increasingly prominent role across various domains. In such a scenario, artificial intelligence could become a central figure in everyday life—robots capable of thinking and acting in ways remarkably similar to humans, performing a wide array of human tasks. These AI-driven robots have the potential to contribute to public welfare, cultural enrichment, and intellectual development in societies. In a hypothetical digital future where robots can create high-quality, artistic works, it raises the question of how well current legal frameworks can retain their relevance. Should they adapt and recognize artificial intelligence as an ‘artificial legal person,’ thereby necessitating the protection of works created by AI? This debate is ongoing in many countries, with varying opinions on whether works generated by artificial intelligence should be granted copyright protection, highlighting the need for legal evolution as software technology advances.
The evolution of AI into a ‘creative’ entity capable of painting, composing music, writing songs, poetry, and stories has sparked numerous discussions, particularly in the realm of copyright law. Central to these discussions is whether the outputs produced by artificial intelligence technology, either directly or with its assistance, can be considered ‘works’ in the artistic sense, falling within the scope of US copyright law. This is tied to the question of whether AI, presently unrecognized as a legal entity by the US legal system, can acquire legal rights through the ‘works’ it produces.
It’s intriguing to note that despite copyright law being an area ripe for development and progress, international consensus on many issues remains elusive. Copyright law fundamentally revolves around art, which is a universal connective thread among people. Hence, it is a highly significant area on the international stage. Particularly in our globalized world, where works created in one country can be accessed simultaneously by many others, the inability to obtain copyright protection in one country while being eligible for it in another is a complex issue that warrants consideration. This complexity is amplified when technological tools are involved and is likely to intensify in the future.
Stephen Thaler has been able to secure copyright protection to some extent in certain jurisdictions, such as the UK, where copyright protection for AI-generated works can be granted to the project developer. However, this limited application is not mirrored in the US. In the United States, even the developers of the AI or the individuals inputting data into the AI program to create an original work do not currently qualify for copyright protection under the existing legal framework. Nevertheless, it remains possible that future legal changes or innovative judicial interpretations may create alternative approaches more aligned with the needs of the digital age, given that the future cannot be envisioned without AI and AI-generated works. A common international standpoint needs to be reached to resolve the intricate disputes and ambiguities.
In the realm of classical law, it is natural and reasonable to contemplate the necessity of a human subject for authorship of a work. Copyright law, after all, is a system that grants protection to ‘original works of authorship,’ safeguarding the ‘fruits of intellectual labor.’ According to this definition, the incapability of artificial intelligence to produce a work in the legal sense aligns practically with the objectives of copyright law, as a non-human entity lacks consciousness, intellectuality, or innate creativity. However, the extent of artificial intelligence technology and its proximity to human consciousness continue to astonish. For instance, systems like DALL.E 2 can generate realistic images, nearly indistinguishable from real photographs, using AI. Similarly, with ChatGPT, artificial intelligence can craft written narratives by processing and utilizing provided data. In light of these advancements, categorizing works produced by artificial intelligence as non-copyrightable is no longer as straightforward as it once seemed. Even though they are not human-made, these works bear striking resemblance to those created by humans. Many reasonable individuals would consider these works, regardless of their producer, as copyrightable when evaluated based on their distinct characteristics and qualities. However, they currently lack copyright protection solely because they were created by a non-human entity. This prompts a deeper exploration and discussion regarding whether these works could indeed be protected by copyright, should the law not stipulate that the work must be exclusively human-created for copyrightability.