CEU Electronic Theses and Dissertations, 2024
| Author | Singay, Muhammad |
|---|---|
| Title | AI Revolution on Trial: Protection of Intellectual Property Rights against Generative Artificial Intelligence in the USA and UK |
| Summary | The advent of Generative Artificial Intelligence (‘Gen-AI’) heralds a transformative era with the potential to reshape human society as we know it. In the global race for AI dominance, prominent programs like Chat GPT, Dall-E, and Stability Diffusion showcase the power of this technology. In the midst of this technological breakthrough, countries find themselves in a tug of war between their desire to dominate Gen-AI industries and to safeguard the Intellectual Property rights of the original creators. The Gen-AI programs are trained on extensive datasets which includes writings, paintings, audios, videos and photographs, referred to as training datasets. These training datasets are primarily compiled through unauthorized data scraping from online sources which may also include copyrighted materials, raising ethical and legal concerns. This thesis specifically explores the legal and economic implications of the usage of copyrighted materials by the Gen-AI companies to train their models without the creator’s permission or compensation. The paper aims to strike a balance between the protection of the rights of copyright holders and fostering innovation through Gen-AI development. This thesis initiates with the exploration of the legal and economic rationale behind copyright protection. The thesis further examines what are the Gen-AI models, how they are trained and how their training process infringes the copyright holders. The thesis also shed light on unauthorized training of Gen-AI models assessing its legal, ethical and economic impacts on creators, users, Gen-AI companies and on countries in general. This thesis then delves deeper into the potential legal defenses for Gen-AI companies against the claims of copyright infringement, particularly the principle of ‘fair use’, ‘fair dealing’, ‘text and data mining exceptions’ and its interpretation under the copyright laws of the United States of America and the United Kingdom. Furthermore, the thesis explores the legal and economic dilemmas faced by the nations like the United States of America and the United Kingdom which are competing to be the forebears of the Gen-AI revolution. The training of Gen-AI programs on the datasets containing copyrighted materials falls under the category of copyright infringement under the laws of the United States of America and the United Kingdom. However, legislation of United Kingdom is planning to revise its copyright laws to facilitate the growth of Gen-AI companies despite of facing serious criticism from the copyright holders. On the other hand, the existing laws of the United States of America are vague on whether using copyrighted materials to train the Gen-AI programs falls under the existing exceptions of copyrights. There is extensive litigation going on the same question in courts of both jurisdictions. The last part of the thesis consists of the analysis of the potential solutions to balance between copyright protection and innovation in the domain of Artificial Intelligence including establishment of clearer legal frameworks and guidelines, collective licensing and royalties, implementation of technological solutions and international cooperation for protection of copyright. The final chapter synthesizes the discussion, draws conclusions and outlines a predictable way forward. |
| Supervisor | Botta,Marco |
| Department | Legal Studies LLM |
| Full text | https://www.etd.ceu.edu/2024/singay_ali.pdf |
Visit the CEU Library.
© 2007-2021, Central European University