On March 16, 2026, the Delaware Court of Chancery issued an opinion finding that the CEO of a major gaming company followed guidance from ChatGPT (referred to here as the “AI Tool“”) to breach a $500 million acquisition agreement. This dispute joins a growing list of cases in which chatbot records have entered the courtroom as evidence against users
Background on the Case
In 2021, Krafton, Inc., a South Korean video game publisher (the “Acquirer”), acquired Unknown Worlds Entertainment (the “Subsidiary”) in a $500 million acquisition. The acquisition terms promised an additional $250 million earnout if revenue targets were hit through December 2025. To hit these targets, the Subsidiary’s pre-sale leadership team remained in control and could only be fired for “cause” – with “cause” defined as the commission of a felony, gross misconduct, disclosure of confidential information, or an “intentional act of fraud or dishonesty.”[1]
According to the Court, as forecasts revealed that the Subsidiary would likely meet its targets, the Acquirer’s CEO searched for ways to avoid the earnout. When his head of corporate development informed him the Subsidiary’s leadership had not committed misconduct–and the earnout couldn’t be avoided –he sought feedback from the AI Tool. Over the next month, the Court found, he “followed most of [the AI Tool]’s recommendations,”[2] leading him to prevent the Subsidiary from launching a product, so that its end-of-year sales fell short of the earnout targets. He then used this to fire the Subsidiary’s key leaders.
Bringing Chatbot Logs Into the Courtroom
Critically, the AI Tool maintained a record of its discussions with the Company’s CEO—and those records featured prominently in the Court’s opinion. The Court included a block quote directly from the CEO’s chats with the AI Tool, outlining strategies that the AI Tool provided the CEO. The Court highlighted three of these strategies, then overlaid each with the CEO’s conduct, including the CEO’s decision to lock down video game distribution channels to prevent the Subsidiary from launching its upcoming product. The AI chat logs thus served as a virtual “witness” that helped “testify” to the CEO’s plans and intent – leading the Court to ultimately find that the Company had breached its contract and issue an order to reinstate the Subsidiary’s CEO and extend its sales timeline by over 8 months.[3] Meanwhile, the litigation continues to determine whether the Acquirer owes any damages.
Key Takeaways and Next Steps
This decision comes at a time when lower courts have begun to grapple with whether and to what extent AI chatbot records can act as evidence in discovery and trial. For example, in the criminal context, the Southern District of New York recently held in United States v. Heppner that documents generated by a public AI platform were not protected by attorney work product privileges, in part because the AI platform’s terms permitted disclosure to government authorities. In this context, the Chancery Court’s opinion underscores that AI chatbot interactions may be discoverable and admissible.
For companies, the opinion offers a pointed lesson: chat logs and other outputs from AI tools may be discoverable in litigation. Depending on what is “discussed” with AI, these materials may be key evidence. Here, the AI-generated paper trail was central to demonstrating that the Company’s actions constituted a deliberate breach of contract rather than a good-faith business judgment.
This comes at a time when businesses are facing increasing demand to roll out AI tools to their workforce to improve productivity and efficiency. While these gains can be achieved, companies would remain well-advised to think through the policies, guidelines, and training they provide to employees governing the use of generative AI tools, so that use is consistent with managing going-forward litigation risk.
For more information, Alston & Bird’s Privacy, Cybersecurity, and Data Strategy Team has extensive experience advising clients on legal issues related to emergent AI products, including U.S. and international AI legislation. Please contact us and sign up for alerts at AlstonPrivacy.com
[1] Fortis Advisors, LLC v. Krafton, Inc., C.A. 2025-0805-LWW, at 15 (Del. Ch. Mar. 9, 2026).
[2] Id. at 35.
[3] Id. at 87–89.
