AI, Privilege, and Discovery: What Recent Cases Mean for Users
Two recent decisions from the Southern District of New York and the Delaware Court of Chancery highlight a growing legal reality: interactions with generative AI tools are not shielded by traditional legal protections. Courts are increasingly recognizing AI chat sessions as third-party communications that are fully discoverable in litigation and, in some cases, powerful evidence of a party’s intent, strategy, and state of mind.
Key Takeaways
AI chat logs are not privileged. Courts view AI platforms as third parties, meaning communications are generally not protected by attorney-client privilege or work-product doctrine and may be discoverable.
AI-generated content can be used as evidence of intent. As seen in Fortis Advisors, courts may rely on AI chat logs to evaluate a party’s decision-making, planning, and good (or bad) faith.
AI tools are not a substitute for legal counsel. AI tools lack the context, judgment, and confidentiality required for legal advice and can create risk if relied upon improperly.
Post hoc sharing with counsel does not restore privilege. Communications that are not privileged when created (e.g., AI chats) do not become privileged simply because they are later shared with an attorney.
In-Depth
U.S. v. Heppner
On February 17, 2026, in the case of U.S. v. Heppner, the U.S. District Court for the Southern District of New York ruled that a criminal defendant’s conversations with an AI chatbot about his case could not be protected by attorney-client privilege or the work-product doctrine. The District Court found the defendant had used the AI to analyze legal issues and assist with communications, but sought to withhold the chats from prosecutors under an assertion of privilege. The court found no privilege applied because the AI platform was a third party, not a lawyer, and the communications were not made in confidence. As a result, the government was allowed to obtain and use the chats in the case.
Fortis Advisors LLC v. Krafton Inc.
In a March 16, 2026, decision in Fortis Advisors LLC v. Krafton Inc., the Delaware Court of Chancery adjudicated claims arising from the sale of a video game studio, Unknown Worlds, where Fortis alleged Krafton acted in bad faith to avoid a $250 million earnout. In enforcing the merger agreement and granting specific performance, the court cited evidence that Krafton’s CEO had consulted ChatGPT extensively, relying on the CEO’s chat logs with ChatGPT as evidence of his intent and planning in orchestrating the takeover.
