Key Takeaways

• Courts are increasingly treating AI chatbot interactions as discoverable electronic records, with recent decisions confirming they fall under existing discovery rules.
• This classification means AI chats carry the same legal risks and obligations as other electronically stored information in litigation, and gaps in preserving these communications can expose companies to sanctions or adverse inferences.
• Companies should consider updating their information governance and litigation-hold practices to account for AI tools used across the organization.

As generative AI tools become part of everyday business workflows, courts are treating chatbot interactions the same way they treat emails, text messages and internal messaging platforms. Recent court decisions confirm that AI chats are not a special category of communication. Rather, they are electronic records that may be discoverable — and may need to be preserved once litigation is reasonably anticipated. In this alert, we outline how courts are approaching AI chats and what companies should do to address preservation obligations.

AI Chats: Existing Rules Apply

Courts are beginning to address AI-related communications in litigation, and the early takeaway is straightforward: existing rules apply. In Fortis Advisors LLC v. Krafton, Inc., the Delaware Court of Chancery repeatedly cited to a CEO’s ChatGPT interactions in its decision, treating them like any other internal business communication like email or Slack.

Other decisions, including Warner v. Gilbarco Inc. and United States v. Heppner, addressed the application of various privileges, but they both upheld the premise that AI chats are potentially discoverable. Together, these cases reinforce that communications with AI tools are subject to the same discovery framework as more familiar forms of communication.

The practical implication is clear: AI chats are another form of electronically stored information, much like email, instant messages and text messages.

Preservation Obligations: Don’t Overlook AI Chats

Because AI chats are discoverable, companies should consider preservation obligations for the chats. The longstanding rule has not changed: once a company reasonably anticipates litigation, it must take reasonable steps to preserve relevant information. For decades, courts have held this preservation obligation applies to electronically stored information, and chatbot interactions are likely no different.

This creates several challenges for companies, as AI chats may not be centrally stored or retained in the same way as email or enterprise messaging systems. Employees may also use personal accounts or non-enterprise AI tools, making identification and collection more difficult. Some platforms may also have limited retention periods or allow users to delete chat histories, creating risk of spoliation if preservation steps are not taken in time.

Failing to preserve relevant AI communications can expose companies to the same risks that apply to other data sources, including sanctions, adverse inference instructions or other remedies.

Practical Steps for Companies

To mitigate these risks, companies should consider what steps it can take to identify and preserve AI chat conversations. This includes:

• Identifying the AI tools in use across the organization, including both enterprise and individual-use platforms.
• Evaluating retention settings and data accessibility for those tools, including how they work with litigation hold software.
• Updating litigation hold procedures to explicitly include AI-generated communications.
• Instructing employees to preserve relevant AI chats when a hold is issued.
  Coordinating with IT and vendors to ensure that relevant data can be collected if needed.

What This Means for Managing AI-Generated Data

Courts are not creating new rules for AI — they are applying existing ones. Communications with chatbots are likely discoverable, which means those conversations, like other ESI, are subject to preservation obligations when a party anticipates litigation.

Companies should therefore consider how to best approach AI tools and consider doing so with the same discipline applied to traditional ESI sources: understand where data resides, ensure it can be retained, and take prompt action to preserve it when legal obligations arise.

For more information on AI-related discovery and preservation obligations, contact Jay Heidrick or your preferred Polsinelli attorney.