06/18/2026
🤖 is changing how legal teams search, analyze, and review information during . But as becomes more integrated into those workflows, a new question is emerging.
❓Could the prompts used to identify relevant documents become discoverable?
A recent decision in Conservation Law Foundation v. Shell Oil Company suggests that the question may be moving from theory to reality. In what appears to be the first court order addressing the production of AI prompts used to identify supporting documents, a federal court required the disclosure of prompts developed and used by an expert witness as part of her methodology.
💬 In his latest article, Phil Favro wrote, "Parties can be certain that Conservation Law Foundation will become a staple citation in negotiations and motion practice over the discovery of AI prompts, especially prompts that parties used to identify relevant information for production in discovery."
While the ruling is limited to expert discovery, its implications extend much further.
The decision raises important questions about:
✅ Relevance and proportionality
✅ Attorney-client privilege and work product protections
✅ Documentation of AI-assisted workflows
✅ Expert methodologies and defensibility
✅ The discoverability of AI prompts and outputs
Favro cautions against drawing overly broad conclusions from the case, noting that the court did not directly address several foundational discovery issues that will likely shape future disputes.
🔍 As organizations increasingly rely on AI to accelerate discovery, investigations, and legal review, understanding how AI-generated work product may be scrutinized is becoming an essential part of governance and risk management.
📌 Read the full analysis to understand why this case could become an important reference point in future discussions about AI, discovery obligations, and defensible legal workflows.
🖇️ https://hubs.ly/Q04lSQnK0
HaystackID®
