Extract from James R. Tyminski and Taskeen Aman’s article “How Gen AI Is Reshaping E-Discovery in IP-Heavy Industries”
As electronic discovery continues to evolve, pharmaceutical and technology companies—particularly those navigating the complexities of patent litigation—face a rapidly changing technological landscape that is increasingly influenced by artificial intelligence (AI) tools. Discovery in U.S. litigation, especially document review, can often be a lengthy and expensive process. This naturally raises the question: why not leverage AI tools more aggressively to streamline the process? Generative AI (gen AI) promises a faster, smarter, and more scalable document review, but not without complications.
Traditionally, e-discovery meant painstaking manual review of documents, emails, and files by teams of attorneys. As data volumes ballooned, predictive tools like computer assisted review (CAR) or technology assisted review (TAR) stepped in to help. These tools use supervised machine learning models that require legal teams to train the software with tagged examples (e.g., responsive versus non-responsive documents), allowing the software to predict relevance across large volumes of documents. While TAR transformed the scope of discovery, particularly in document-heavy patent litigations, it remained a largely reactive tool.
Enter gen AI. This latest generation of AI tools, powered by large language models (LLMs), has the potential to go far beyond TAR.
