AI Chatbot Logs Now Treated as Discoverable Evidence in Court
Legal experts warn that prompts and AI transcripts are being used to prove bad faith and negligence in US and UK litigation.
Corporate interactions with generative AI are no longer private, as courts in the U.S. and UK increasingly treat chatbot prompts and logs as discoverable evidence. Legal experts from firms including Blake, Cassels & Graydon LLP and Spencer Fane warn that these digital trails are being used to establish strategic intent, professional negligence, and bad faith.
Recent rulings demonstrate that AI chat logs are being categorized as electronically stored information (ESI), making them subject to the same production rules as emails. In Fortis Advisors, LLC v. Krafton, Inc., the Delaware Court of Chancery utilized a CEO's ChatGPT queries and logs to establish motive and pretext in a breach of contract dispute over earnout obligations, treating the prompts as evidence of bad faith. Similarly, in Conservation Law Foundation, Inc. v. Shell Oil Co., a U.S. District Court in Connecticut ordered the disclosure of all AI prompts used by an expert witness, ruling that the prompts were part of the expert's methodology and therefore discoverable.
The Shift in Digital Discovery
As generative AI becomes a staple of business workflows for drafting memos and brainstorming strategy, many organizations have operated under the mistaken assumption that these interactions are ephemeral or off-the-record. However, the legal framework for ESI has evolved to encompass AI transcripts. This shift means that any prompt entered into a chatbot can potentially be produced during the discovery phase of a lawsuit.
This trend extends to professional liability. In the UK High Court case Anthony Malcolm Cork & Anor v. Mark Smith, the court addressed the submission of AI-generated material that contained hallucinated statutory provisions. The court found that the lawyer involved had failed to verify the information, despite the chatbot providing a warning to do so, highlighting how AI logs can be used to prove professional negligence.
Corporate Liability Risks
This development creates a significant new liability risk for corporations and legal professionals. If litigation hold templates are not updated to explicitly include AI prompts and transcripts, companies may face sanctions for the spoliation of evidence. The permanent nature of these logs means that a single strategic brainstorming session with an AI can later be used to establish a party's state of mind or intent in court.
Spencer Fane noted that "your chatbot may not be your friend when litigation arrives," while Emily Hazlett and Hannah Cinel of Blake, Cassels & Graydon LLP advised professionals not to communicate with an AI chatbot in a way they would not want presented before a judge.
What to Watch
Legal departments must now treat AI usage as a formal record-keeping activity. The primary focus moving forward will be the integration of AI logs into standard data retention and discovery policies. While the core principle of discoverability is now established through case law, the specific boundaries of what constitutes a "protected" interaction versus a discoverable business record remain a key area of evolving litigation.