Recently, we were all reminded that any type of internal communications – including AI prompts! – are discoverable and can become the “smoking gun” in a lawsuit. In a $250 million M&A earnout dispute, as reflected in this court’s post-trial decision, the buyer was found by the Delaware Court of Chancery to have intentionally undermined the earnout and prevented the seller from achieving it.
A key piece of evidence came from the CEO’s AI chatbot prompts, which exposed his true intent behind terminating seller executives and assuming operational control of the acquired company. Acting on the AI chat’s suggestions, the CEO created an internal initiative called “Project X,” aimed at either renegotiating the earnout or executing a full takeover of the target company (as noted on page 33 of the court’s decision).
Vice Chancellor Will quoted the CEO’s AI chat conversation in quite a bit of detail. As noted on page 35 of the court’s decision, included in that was the AI chatbot’s detailed “Response Strategy to a No-Deal Scenario,” which included a “pressure or leverage package” of:
- Negotiation talking points
- Preemptive framing to attract customer trust
- Steps to lock down the target’s product rights
- Guidance on preparing materials for a potential legal defense
- A “two-handed strategy” combining hardball legal and financial pressure with softer support and incentive approaches
The CEO admitted during trial that he had deleted AI chat logs, raising additional concerns about intent and evidence preservation.
Although the opinion doesn’t seem to establish a new legal standard governing AI use, it does provide a compliance reminder that interacting with AI models isn’t necessarily privileged and may be subject to discovery.
Authored by
Broc Romanek