Litigation Minute: Is AI-Generated Content Discoverable? What Companies Need to Know in 2026
Recent court decisions indicate that AI-generated data, including prompts and outputs, is discoverable but must adhere to relevance and proportionality standards.Summary
The rapid adoption of AI tools, particularly generative AI (GenAI), is creating new challenges for e-discovery. Courts are now addressing whether data related to GenAI – prompts, outputs, and activity logs – falls under traditional discovery obligations. Recent rulings, such as *In re OpenAI, Inc.*, confirm that relevant GenAI data *is* discoverable, mirroring the treatment of other ESI. However, proportionality remains a key consideration; while large volumes of data may be discoverable if justified by the case's needs, courts will scrutinize the burden of production.
Companies must proactively prepare for GenAI-related discovery by identifying custodians using these tools, preserving potentially relevant data (including activity logs), and negotiating the scope of discovery early on. This includes addressing confidentiality concerns through protective orders and anonymization. Updating information governance policies to incorporate GenAI data into ESI inventories and retention policies is also crucial.
Ultimately, courts are applying established discovery principles to this new data type. Companies should work closely with e-discovery specialists to minimize burden, manage privacy concerns, and ensure a defensible approach to GenAI data discovery, recognizing that data central to a dispute is likely discoverable, but proportionality will be a limiting factor.
(Source：National Law Review)
