Legal Considerations for Generative AI in Games

On January 5, 2021, OpenAI launched DALL-E, a neural network trained on 250 million captioned images collected from the internet that could generate images based on natural language prompts. Since then, OpenAI has introduced DALL-E 2, which generates more realistic, high-resolution images than its predecessor, and ChatGPT, a conversational language model of artificial intelligence capable of answering follow-up questions, admitting its mistakes, challenging incorrect premises, and rejecting inappropriate requests. A flurry of other text-to-image generators have also emerged, such as Stability AI Ltd.’s (“Stability AI”) Stable Diffusion and Midjourney, and in only two years, artificial intelligence that can generate novel content (“Generative AI”) rather than simply analyzing or acting on existing data, has rapidly become the technology du jour.

January 13, 2023, however, a group of artists filed a class action lawsuit against Stability AI, Midjourney, and Deviant Art based on the companies’ text-to-image Generative AI tools, which the complainants allege have infringed on the rights of thousands of artists (the “Artists’ Suit”). Although Generative AI has utility to creatives, including game developers, seeking to use Generative AI tools to save money and time while expanding content, the Artists’ Suit highlights tension between existing law governing creators’ rights and the evolving field of Generative AI. This legal update, drafted with the assistance of OpenAI’s ChatGPT, outlines the questions of ownership and copyright infringement raised by Generative AI in the context of video games, and compares the terms of service of several popular Generative AI tools in appreciation of the fact that some questions are addressed by contract rather than copyright law.