Ready (Or Not) Player One: AI Forces Intellectual Property to Gaming’s Forefront

Artificial intelligence, despite its potential to transform business sectors, scientific discovery, and human interaction, is not without its detractors. One notable area of contention lies within the video game industry, as developers must increasingly grapple with complex legal challenges arising from the integration of the technology in their titles.

Video game developers have counted on AI as a cornerstone since the industry’s genesis in the 1950s. In gaming, AI serves a unique purpose – it is designed to create logically responsive, adaptive, or intelligent behaviors predominantly in non-player characters (NPCs) to mirror human-like intelligence. This is markedly different from AI based on machine learning or decision-making capabilities.

In the early days of arcade video games, developers employed AI primarily to enhance gameplay through graduated difficulty levels, distinct movement patterns, and in-game events triggered by the player’s actions. AI does not aim to create an artificial persona corresponding to an NPC. Instead, it serves as a tool to enrich the player’s experience by making the game more interactive and engaging.

In the modern era, one of the most significant advantages AI brings to gaming is the ability to perform extensive data mining on player behaviors. This provides invaluable insights into how players interact with the games, identifies their most popular features, and pinpoints the factors that cause games to fall out of favor. This information enables video game designers and studios to fine-tune gameplay elements, identify potential monetization opportunities, and incorporate winning attributes into new games.

However, integrating generative AI into video games has created a tangle of intellectual property rights issues. As AI becomes more sophisticated, autonomous, and creative, questions about ownership, authorship, and responsibility become more nuanced. These challenges necessitate a thorough understanding of both AI technology and IP law, presenting a formidable task for video game developers.

Intellectual Property Issues

AI’s growing role in the entertainment industry has exposed existing IP laws’ inadequacy in regulating the technology’s rapid evolution. As AI becomes more powerful, laws must be created that define authorship, assign responsibility guard against plagiarism, and fairly apply IP law.

IP Ownership
Generative AI introduces unique scenarios that did not exist when current IP laws were adopted. In most jurisdictions, copyright laws extend automatic protection to human authors of original works. However, this protection does not seem to encompass AI-generated works.

According to the AI Registration Guidance issued by the US Copyright Office, copyright protection in the US necessitates human authorship as the term is used in both the Constitution and the US Copyright Act, and excludes non-human entities. Furthermore, a US district judge ruled in April that art created solely by AI cannot be granted copyright protection. So, if copyright holders must be human, and the humans who program or prompt AI for its output are not allowed to register a copyright for it, what protection does AI-generated content have?

The intricate issue of ownership becomes more convoluted when AI systems are trained using a blend of human-created, copyrighted material and public domain data. In some instances, the legal framework acknowledges human involvement in training AI systems as a pivotal factor in determining ownership. If a human offers significant creative input or makes informed decisions that influence the AI-generated output, they could be regarded as the rightful owner of the resulting work.

Interestingly, the US law covering copyright ownership over generative AI works in the US differs from several other countries. Notably, Hong Kong, India, Ireland, New Zealand, and the UK explicitly grant authorship rights to programmers, that is, the humans involved in developing or training the AI system. This divergence in global approaches highlights the need for comprehensive, internationally harmonized regulations addressing the complex IP issues arising from AI technologies.

IP Licensing
The use of copyrighted material in training these AI models has elicited claims from copyright holders who argue that platform developers should have obtained permission before repurposing their content.

Getty Images has sued Stability AI, alleging Stability’s diffusion art generator copied millions of images and their affiliated metadata and generated some “bizarre or grotesque” output that included Getty’s watermark, damaging its reputation for quality.

Different jurisdictions are addressing these concerns in varying ways. The United Kingdom, for instance, is streamlining the process for AI developers by easing the rules protecting copyrighted material. This allows AI developers to utilize such content for training AI systems without needing explicit permission from the rights holder. This relaxation of rules forms part of the UK’s strategy to position itself as a prime location for AI innovation and research.

The United States and the European Union have not made similar provisions, maintaining stricter control over copyrighted material. The EU’s newly drafted AI Act mandates developers to publish an exhaustive list of all copyrighted material used to train AI models. This could potentially trigger a wave of lawsuits from copyright holders who were previously unaware that their work had been used for AI training.