The first recorded idea of “a machine that thinks” dates back to ancient Greece, but it was not until 1956 that John McCarthy would first coin the term “artificial intelligence” (or “AI”) during the first-ever AI conference at Dartmouth College. That same year, Allen Newell, J.C. Shaw, and Herbert Simon created the Logic Theorist, the world’s first ever running AI software program. For decades, companies like IBM would invest in artificial intelligence research, even creating artificial intelligence capable of beating world-champion Go players. In 2013, the first generative artificial intelligence program capable of generating realistic images and speech was created. Since then, more generative artificial intelligence has emerged, from GPT-3 to DALL-E 2, showing the impressive creative potential of artificial intelligence. Now, around the world, companies are investing more and more money into artificial intelligence (AI) platforms. By 2026, global spending on artificial intelligence is expected to exceed $301 billion.

     With the advent of generative AI, questions arise concerning how these technological advancements interact with copyright law. Given the elements of copyrightability discussed in Section II of this paper, past scholarship has focused on whether AI itself should be able to hold copyrights for its created works, with authors coming to conclusions based on implications of the decision and the current state of law. This especially becomes relevant as AI becomes more utilized. While authors debate whether or not AI-generated works should be copyrightable and the factors to consider, many authors have not looked to what other countries have done and the implications of their decisions to help guide the United States approach.

     In this paper, I conduct a comparative analysis of global approaches to both copyright and patent law regarding AI-generated works. Finally, I conclude the paper with recommendations about how to address this question and possible solutions.10 Section I discusses how artificial intelligence creates or assists in creating work, as the mechanisms used can help elucidate whether the resulting works should be copyrightable and by who. Section II discusses copyright law in the United States, while Section III touches on how copyright law and generative artificial intelligence intersect. Section IV explores how the United States Copyright Office has dealt with generative AI output copyrightability questions, and Section V explores how China and the European Union have addressed the issue, with hopes that their approaches will elucidate the approach that the United States should take. Finally, Section VI concludes the paper by arguing for allowing copyright of AI-generated outputs as long as there is sufficient human interference.