AI records are becoming increasingly important evidence in litigation. When a CEO turned to AI to sidestep making a $250 million payout, he did not except the logs of the conversation to end up in court, let alone that they would be crucial evidence in the case. From transcribing meetings to aiding in business decisions, AI platforms are becoming an increasingly important part of workplaces. AI can be a powerful tool with results such as greater efficiency and new creative possibilities. However, AI usage can also generate additional risks, particularly when it comes to the generation of records and their potential inclusion in litigation.
Key AI Records Litigation Takeaways
- Internal AI chat logs, prompts, and automated meeting transcripts are legally discoverable and subject to heavy scrutiny by courts.
- Attorney-Client Privilege generally does not protect conversations with AI tools, even if legal advice is being sought.
- Rulings on the Work Product Doctrine vary sharply by state, often depending on the state’s public privacy policies.
- To mitigate litigation risks, organizations must map AI usage and align outputs with corporate retention schedules.
Potentially Risky AI Records
Some AI-related records concerns are easily anticipated such as reliance on incorrect information and privacy vulnerabilities. Far fewer organizations scrutinize the records that AI usage produces.
Even using AI for routine tasks can cause concerns to crop up. AI-generated meeting transcripts and summaries can transform candid or offhand comments into lasting records.
The records produced by AI chat and conversation tools warrant particular attention. While employees might be using AI chat sessions as casual brainstorming tools, they may not realize that the prompts, outputs, and logs generated are often retained within these tools. Conversely, some tools apply auto-delete policies meaning that records can be erased regardless of their importance.
Whether unknowingly retained or deleted, these records increasingly factor into litigation.
Chat Log Case Study
In Fortis Advisors LLC v. Krafton, Inc., former stockholders of video game studio Unknown Worlds sought a payout dependent on post-acquisition revenue. The agreement with the acquirer Krafton allowed the studio’s founders and CEO to retain operational control and ensured they could only be dismissed in strictly defined circumstances.
Several months before the payout period concluded, Krafton terminated the founders and CEO claiming that their plans to release a new game would damage the company. Discovery later revealed a different version of the story: the Krafton CEO received projections that the game’s launch would likely trigger the payout. His next step was to consult AI for advice on how to avoid that outcome. The chatbot responded with a strategy, and the CEO took steps aligning with these recommendations, which led to terminating the founders.
Not only were the chat logs deemed discoverable, but the court relied on them extensively to find that the CEO’s reasons for the terminations were pretextual. AI logs are subject to scrutiny in litigation, and courts will not necessarily dismiss them as informal brainstorming. These records can become key evidence.
AI Records Litigation in Existing Legal Frameworks
This case is part of a growing wave testing how AI records fit into established evidentiary rules. Two doctrines are getting the most attention:
Attorney-Client Privilege: Generally Inapplicable
Multiple courts have held that AI tools are not attorneys, and AI conversations are not confidential. The privilege will not apply to conversations even if users seek legal advice from AI tools.
Work Product Doctrine: Murkier Waters
There is less consistency for the Work Product Doctrine:
- Delaware: AI records generated in the ordinary course of business were not protected regardless of forthcoming legislation.
- Michigan: Defendants attempted to compel a self-represented plaintiff to produce AI records related to the suit. The judge held that the doctrine applied, since the records were generated in anticipation of litigation and AI platforms are tools, not third parties, so using them is not a disclosure to outside persons.
- New York: Doctrine did not apply because the defendant’s use of a publicly available AI tool with privacy policies disclosing that prompts can be used to train models and shared with third parties negated any reasonable expectation of privacy.
As a final complication, the normally good hygiene practice of deleting records routinely can become evidence of inappropriate destruction of records in the face of litigation.
Next Steps for Businesses
As demand for AI records management increases, organizations should get ahead of the issue:
- Map current usage: Understand how employees are using AI – what types of inputs and outputs are being generated and managing the distinction between records and non-records
- Training: Educate employees on what AI captures and how to best leverage technology in areas such as auto-deletion to avoid manual processes
- Anticipate uncertainty: Prepare for unpredictable responses to AI records as the law catches up with technology
- Map records to retention schedules: Address AI records proactively rather than after litigation forces the question and base mapping on business processes
- Litigation holds: Understand the procedures for overriding in place retention and auto-delete policies when litigation is anticipated
As Fortis makes clear, the question isn’t whether AI records will appear in litigation, but whether companies will be prepared when they do so.
AI Records Litigation FAQ
Are AI chat logs admissible in court?
Yes. As seen in Fortis Advisors LLC v. Krafton, Inc., courts treat retained AI chat logs, prompts, and system outputs as discoverable business records that can be used as primary evidence.
Does Attorney-Client Privilege cover AI legal tools?
No. Multiple court rulings indicate that AI platforms are not legal counsel, and sharing data with them lacks the required confidentiality to establish privilege.
Disclaimer: The purpose of this post is to provide general education on records management and information governance topics. The statements are informational only and do not constitute legal advice. If you have specific questions regarding the application of the law to your business activities, you should seek the advice of your legal counsel.