Courts Declare AI Chat Logs Discoverable, Raising Preservation Duties for Companies

Context Generative AI tools have moved from experimental labs into daily corporate workflows. Executives draft strategies, legal teams seek advice, and sales staff generate pitches using chatbots such as ChatGPT. As these interactions become routine, courts are deciding how existing litigation rules apply.

Key Facts - The Delaware Court of Chancery, in *Fortis Advisors v. Krafton*, ruled that a CEO’s ChatGPT messages are discoverable electronic records, equating them with email or Slack. - Existing discovery statutes already require parties to produce electronically stored information (ESI) when requested. The court’s decision confirms that AI chats fall within that definition. - Failure to preserve relevant AI chats can trigger the same sanctions applied to other lost data, including adverse inference rulings that allow judges to assume the missing information would have been unfavorable to the non‑preserving party.

What It Means Companies must now treat AI chat logs as a distinct data source in their information governance programs. The first step is to inventory every AI platform used across the enterprise, from corporate‑wide tools to personal accounts accessed on employee devices. Next, firms should audit retention settings; many chatbot services automatically delete histories after a short period or allow users to erase messages, creating a risk of spoliation if preservation steps are delayed.

Litigation‑hold procedures need explicit language covering AI‑generated communications. When a hold is issued, legal and IT teams must instruct employees to retain any relevant chatbot exchanges and to refrain from deleting them. Coordination with vendors is essential to ensure that chat logs can be exported and placed on hold, especially for cloud‑based services that store data outside the company’s direct control.

The practical impact mirrors that of email or instant‑messaging preservation: early identification, swift preservation, and thorough documentation. Firms that ignore AI chat logs expose themselves to the same penalties courts have imposed for lost emails, including monetary fines and adverse inference instructions that can sway a case’s outcome.

Looking Ahead Watch for additional rulings that clarify privilege claims over AI‑generated content and for industry guidelines on standardizing AI data retention. Companies that embed AI preservation into their legal‑tech stack now will avoid costly surprises as litigation involving chatbot evidence expands.