In the rapidly developing landscape of legal technology, courts are increasingly grappling with the implications of generative AI in legal practice. The utilisation of AI by lawyers is progressing at a pace that presents challenges to traditional legal frameworks, particularly around the protection of work product and the boundaries of discovery.
A notable case shedding light on this issue is Tremblay v. OpenAI, Inc. in the Northern District of California (Case No. 23-cv-03223-AMO). The case, which involved copyright owners accusing OpenAI of improperly using their copyrighted creations, provided an early judicial analysis of how work product protection can be applied in the context of AI.
In this case, OpenAI requested all of the plaintiffs’ test results, which included prompts and outputs that did not necessarily reproduce or summarise the plaintiffs’ works, nor support their claims. The request for these ‘negative results’ was a point of contention. Initially, a magistrate judge ordered the plaintiffs to produce both negative and positive test results. However, U.S. District Judge Araceli Martintez-Olguin later reversed this decision.
Judge Martintez-Olguin's ruling was significant as it underscored that the prompts crafted by the plaintiffs' counsel were considered to contain mental impressions and strategic insights, thus rendering them part of the lawyer’s work product. The ruling indicated that while disclosure of prompts that supported the plaintiffs' claims waived their work product protection, this waiver did not extend to other prompts not disclosed in the complaint.
This case highlights the complexities surrounding the use of AI in the legal field. Lawyers frequently utilise AI for tasks such as drafting and filing documents, which raises new challenges around disclosure and the protection of confidential legal strategies. Many courts now require lawyers to disclose their use of AI in preparing filings and are particularly vigilant against AI 'hallucinations'—outputs that are inaccurate or nonsensical.
Moreover, the ruling aligns with the principles outlined in the Sporck doctrine, which traditionally protects a lawyer’s strategic selection of materials, like witness interviews or document compilations used during case preparation, from discovery by adversaries.
As the integration of AI in legal processes continues to grow, more court cases are likely to emerge, requiring further interpretation of work product protections in this context. The legal community is observing closely as these precedents shape the future interaction between technology and law, balancing innovation with the preservation of established legal principles.
Source: Noah Wire Services