A Video Game Lawsuit Shows How Discovery Into AI Chat Logs Is Changing The Game

A recent Delaware Court of Chancery opinion arising from the acquisition of a video game studio carries an important lesson far beyond M&A litigation: AI chatbot exchanges are now entering the discovery record and affecting case outcomes. 

In Fortis Advisors LLC v. Krafton, Inc., the dispute was a familiar one: whether, years after an acquisition, former stockholders of the acquired company were entitled to an additional earnout payment based on post-closing performance. What makes the case unique—at least for now—is the evidence the Court found especially persuasive: the buyer CEO’s chatbot prompts and the AI-generated responses.

For litigators across all practice areas and the clients they counsel, the takeaway is immediate: AI chat logs have entered the e-discovery stream and may be consequential evidence. That development will not stay confined to nine-figure Delaware corporate disputes. It is coming to civil litigation broadly, and lawyers and their clients need to be thinking about it now.

Background

The plaintiff represents the former stockholders of Unknown Worlds Entertainment, the studio behind video game Subnautica. In 2021, Krafton acquired Unknown Worlds for a $500 million upfront payment, plus up to $250 million in contingent earnout payments tied to revenue through 2025. The agreement preserved operational control for Unknown Worlds’ founders and CEO, who Krafton could only remove for narrowly defined cause.

In July 2025—weeks before the anticipated release of Subnautica 2 and months before the earnout period ended—Krafton terminated the founders and CEO citing concerns about their supposed “intention to proceed with the premature release of Subnautica 2,” which Krafton claimed would damage the franchise. Krafton took over control of Unknown Worlds’ operations and delayed the game’s launch.

Unknown World’s former stockholders sued, claiming Krafton’s actions were a breach of the agreement.

Discovery revealed that, just weeks before the terminations, Krafton’s CEO received internal projections showing that a successful launch of Subnautica 2 in August 2025—as Unknown Worlds was planning—would likely trigger most or all of the $250 million earnout. Concerned that Krafton had overpaid, the CEO turned to an AI chatbot for advice on how to avoid the obligation.

After the AI chatbot told the CEO the earnout would be “difficult to cancel,” he asked it for a path forward. The chatbot suggested forming an internal task with a mandate to either negotiate a deal on the earnout or execute a “takeover” of Unknown Worlds. It also generated a “Response Strategy to a ‘No-Deal Scenario,” including a “pressure and leverage package” and an “implementation roadmap.”

Over the following weeks, the CEO and his team took steps that closely tracked those recommendations, culminating in the July terminations. 

In light of this evidence, Vice Chancellor Will found that Krafton’s stated justifications were pretextual and that its actions breached the agreement. The Court reinstated the former CEO with full operational authority and extended the earnout period to account for the time lost during the wrongful termination. 

What This Case Means for Litigants

The CEO’s exchanges with the chatbot did not just add color to the narrative — they were central to the court’s finding that Krafton’s stated justifications for the terminations were pretextual and that its true motive was financial.

As clients increasingly rely on AI tools in the ordinary course of business, those interactions are likely to surface in litigation. And when they do, they may provide a uniquely direct record of strategy, intent, and contemporaneous thinking. The question now is how to manage these risks. Some practical steps lawyers and their clients can take include:

• Treat AI-assisted strategy sessions with the same intentionality brought to other confidential documents. That means understanding when a client’s use of these tools may be protected from disclosure, and when it may not be.AI-assisted exchanges may or may not be privileged depending on who uses the tool, for what purpose, and on what platform.
• Update AI-use and retention policies. Identify which tools are being used, what data is retained and for how long, and who has access.
• Include AI content in litigation holds. Preservation should include chatbot prompts, outputs, and related materials.
• Prepare for discovery and depositions. Counsel should understand how key custodians use AI tools and be ready to address that usage in document discovery and testimony.
• Incorporate AI into offensive discovery strategy. Discovery requests, custodian interviews, and deposition outlines may need to specifically address AI platforms and usage.

Fortis v. Krafton is an early and unusually public example of how AI-generated material can shape the course and outcome of high-stakes litigation. It is unlikely to remain an outlier for long.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.