There has been a lot of push for adopting alternative dispute resolution (ADR) for intellectual property cases. However, is ADR a good idea when it comes to disputes of great social importance such as the recent popular AI-copyright disputes? ADR, despite all its potential benefits for the parties who use it, also has the tendency to come with costs to society, particularly with respect to the development of legal jurisprudence. Unlike trials, ADR processes, including mediation and arbitration, are confidential. In a mediation situation, any evidence produced for the mediation may not be used in a subsequent proceeding involving the same parties, except for evidence that would be discoverable and admissible at trial regardless of the mediation proceeding.
Cases like Feist Publications, Inc. v. Rural Telephone Service Co., where the Supreme Court outlined the criteria for copyright protection’s originality, and Sony Corp. of America v. Universal City Studios, Inc. where the Court helped clarify the “fair use” defense, stand as important legal rulings. The copyright industry would agree that these cases exemplify situations where resorting to mediation would have been regrettable, maybe not for the parties, but for the development of the IP jurisprudence. While these cases may not have saved the parties themselves time and resources, they have allowed for better prediction of outcomes of future cases (and even aided parties who use ADR as this prediction could aid negotiation).
The prominence of high-stakes AI copyright cases has brought attention to the complexities of machine-generated content and its relationship with copyright law. Models like OpenAI’s ChatGPT and Microsoft’s Bing Chat possess the capacity to produce content resembling human-created works, blurring the lines of authorship and ownership. Questions arising from this case include whether the training of AI models with existing copyrighted works is lawful and whether such usage constitutes copyright infringement or qualifies as fair use. In these cases, the choice to opt for settlements risks hindering the establishment of clear legal guidelines and precedents.
With regards to recent AI-generated content-copyright disputes, there has been talk about settlements by parties involved in top-tier cases, including The New York Times’ lawsuit against OpenAI and Microsoft. This consideration of settlements, while aiming for swift closure, inadvertently deprives the broader legal community of crucial precedents that could significantly impact the future resolution of similar disputes.
Settlements (as in alternatives to litigation), while expedient for the involved parties (arguably in the short term in these big AI cases), forego the opportunity to set concrete legal boundaries and precedents essential for navigating the nuances of AI-generated content. The absence of definitive rulings in these cases not only affects immediate disputes but also inhibits the broader community from obtaining essential guidance on applying copyright law to emerging AI technologies. To encourage these kinds of cases to come to court, perhaps expedited hearings could be allowed.
The AI industry should know that they risk perpetuating ambiguity and impeding the development of a robust legal framework crucial for addressing the challenges posed by AI-generated content when they settle critical questions privately. The lack of established precedent impacts not only current disputes but also hampers the evolution of a coherent legal foundation for AI and copyright law. While parties in disputes of societal importance such as these AI cases cannot be forced to go to trial, they should at least be encouraged to consider the impact of privately handling these matters.
 For example, see Jacques De Werra, ‘Can Alternative Dispute Resolution Mechanisms Become the Default Method for Solving International Intellectual Property Disputes’ (2012) 43 Cal W Int’l LJ 39
 Nancy Neal Yeend & Cathy E. Rincon, ‘ADR and Intellectual Property: A Prudent Option’ (1996) 36 IDEA 601