How GenAI Is Reshaping Legal Discovery

EDiscovery Today founder Doug Austin interviewed Cristin Taylor at the recent Relativity Fest on how generative AI (genAI) is reshaping legal discovery. A central concern was defensibility, particularly whether genAI prompts should be discoverable. The question is whether prompts should be protected as privileged work product.

Traylor framed prompts not as simple search terms, but as functional equivalents of review protocols, strategic instructions that guide decision making. Since training materials and review protocols are treated as privileged work product, the same protection should extend to the instructions provided to AI. As Traylor observed, “no one ever asks for your review protocol when you’re training humans.”

From this perspective, the emphasis should shift away from disclosing internal processes and toward demonstrating that outcomes are reliable and reasonable. The cornerstone of defensible AI use then becomes outcome validation, which aligns with long-standing discovery principles. 

Rather than demanding perfection, courts expect reasonableness supported by measurable quality controls. Traylor argued that sharing performance metrics and validation results provides opposing parties with meaningful assurance without exposing privileged strategy. When disputes arise, established discovery practices remain sufficient to address gaps or challenges.

Responsible AI adoption for legal teams begins with ethics, and Traylor noted that existing rules already offer a solid foundation. Integral to that framework are supervision and quality control, which Traylor underscored as essential ethical duties requiring rigorous validation and ongoing oversight.

Traylor highlighted emerging risks, including AI-generated hallucinations in legal filings, underscoring the need for human review. Compared to earlier technology-assisted review (TAR), genAI offers greater transparency by explaining its reasoning, providing insight into why documents are deemed relevant. 

Looking beyond how genAI is reshaping legal discovery today, Traylor sees AI’s greatest value as transforming data into strategic insight. By accelerating knowledge construction, genAI will enable legal teams to move beyond document production toward deeper understanding, better advice, and more informed decision-making.