Navigating the Legal Complexities of Generative AI

The recent surge in popularity of generative AI, largely driven by ChatGPT, has propelled this technology from a niche area of interest to mainstream attention, with significant implications for various industries, including gaming. Unlike previous tech trends like VR/AR and the Metaverse, generative AI boasts immediate and practical use cases, making it an exciting yet challenging field to navigate. As studios experiment with generative AI, largely behind closed doors, they must consider its potential applications in art, music, code, level design, and more, while being mindful of the legal landscape. The rapid adoption of new technologies often catches the law off guard, leading to a period of adaptation as courts and lawmakers grapple with how to regulate them. This is particularly true for generative AI, with each country having its own laws and court decisions at different stages of addressing this phenomenon. A significant challenge is the risk of intellectual property (IP) infringement and ownership issues. The use of generative AI may infringe on copyright, particularly if the training materials were used without a license from the copyright owner. The UK Government's consultation on extending the Text and Data Mining exception to copyright infringement was initially favorable but has since been scrapped, meaning that a license is required to use third-party works for training generative AI under English law. The legal proceedings against Stability AI by Getty for the misuse of library images highlight the kind of disputes that may arise. Questions also remain about the potential liability of the user of generative AI, especially if they use it to substantially copy existing copyright works. The ownership of IP rights in the output of generative AI is another complex issue. In the UK, the Copyright Designs and Patents Act suggests that works with no human author can be protected by copyright, but it is unclear who would be considered the author in the case of generative AI. The person who wrote the code, chose the training materials, or input the prompt could all be considered authors, or it could be a combination of these individuals. In the US, works created by AI are not currently protected by copyright because they lack a human author. However, the law is evolving, and these issues are being considered by the Supreme Court. Given these uncertainties, studios should exercise caution when integrating generative AI outputs into their games. To navigate these challenges, it is recommended to maintain a clear separation between the use of generative AI for inspiration and the final product, keep accurate records of the ideation process, avoid creating a trail of potentially incriminating evidence, consult with in-house legal teams, and stay informed about the rapidly evolving legal landscape. Other potential legal issues include data protection and privacy implications, as well as moral considerations. While generative AI presents significant opportunities for speeding up creative processes, cutting costs, and contributing to the democratization of the industry, it is crucial to approach its use with careful consideration of the legal and ethical implications.