Delaware Court Found Krafton Followed Most of ChatGPT's Recommendations in Campaign that Wrongfully Terminated Unknown Worlds Executives and Seized Operational Control

A Delaware Chancery opinion found that, after being warned that firing Unknown Worlds leaders would not eliminate earnout obligations, Krafton's CEO turned to ChatGPT for a "no-deal" strategy and then followed most of its recommendations. The court tied that strategy to Krafton's lockout of publishing control, unilateral public messaging, wrongful termination of Ted Gill, Charlie Cleveland, and Max McGuire, and seizure of studio control.

Risk classification

• Primary risk domain: 5 Human-Computer Interaction
• Primary risk subdomain: 5.1 Overreliance and unsafe use

The Krafton CEO inappropriately relied on ChatGPT to guide critical legal and corporate strategy during a major business dispute, leading to wrongful executive terminations.

Additional risk subdomains

• 4.3 Fraud, scams, and targeted manipulation: The CEO used ChatGPT to generate a bad-faith strategy to avoid paying a $250 million earnout and execute an improper corporate takeover.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was caused by Krafton's CEO intentionally deciding to use ChatGPT to generate and execute a bad-faith business takeover strategy.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: ChatGPT is a general-purpose chatbot. The report describes a user interacting with a chatbot to receive advice, which falls under the transparency obligations of Limited Risk systems.

AI system and alleged parties

• AI system: None named
• AI purpose: Writing Assistant; Question Answering
• Behaviour type: Assistant
• Alleged developer: OpenAI
• Alleged deployer: Krafton, Changhan (CH) Kim
• Alleged harmed parties: Unknown Worlds Entertainment, Ted Gill, Max McGuire, Former shareholders of Unknown Worlds Entertainment, Charlie Cleveland

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Negligible. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Minor
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: The report describes a dispute over a $250 million earnout and a $500 million acquisition, though the actual financial loss from the wrongful termination and subsequent legal battle is not fully quantified.

Directly caused: The report does not specify a direct liquidated financial loss, but notes Krafton attempted to avoid a $250 million earnout payment and spent resources executing the improper takeover.

Indirectly caused: The wrongful termination of Unknown Worlds leadership likely caused temporary loss of earnings for the executives and legal expenses for both parties, though specific figures are not detailed.

Inferred additional harm: It is highly likely that both Krafton and the terminated executives incurred millions of dollars in legal fees during the Delaware Court of Chancery proceedings, and the studio suffered operational disruptions.

People affected

• Occurrences reported: 1
• People reportedly harmed: 3
• People reportedly exposed: 3

Potential causes

Management

• Prioritizing Cost Savings: Management focused on avoiding the 250 million dollar earnout at all costs.
• Improper Studio Takeover: Executing a hostile takeover of an independent studio based on AI advice.
• Unjustified Firing: Removing studio leadership under false claims to execute the AI plan.

Technology

• Unchecked AI Strategy Generation: ChatGPT provided complex takeover strategies without legal compliance checks.
• Lack of Contextual Safeguards: The AI tool lacked guardrails preventing its use for contract circumvention.

Data Inputs

• One-Sided Prompting: The CEO input biased details about the contract to get desired AI advice.

Human Factors

• Overreliance on AI Advice: The CEO blindly followed ChatGPT recommendations for a major business dispute.
• Fear of Pushover Deal: CEO anxiety over a 250 million dollar earnout led to desperate decisions.

Process and Methods

• Bypassing Legal Consultation: The company substituted professional legal counsel with chatbot advice.
• Flawed Dispute Process: Using automated recommendations to handle complex acquisition agreements.

Regulatory Environment

• Lack of AI Governance Rules: Absence of regulatory standards on using AI for corporate decision-making.

Information quality

• Classification confidence: High
• Reason for confidence: The report clearly outlines the role of ChatGPT in generating the takeover strategy, the specific parties involved, the financial figures of the dispute, and the court's ruling. There is little ambiguity regarding what transpired.
• Alternative interpretations: One could argue the incident is purely a human contract dispute, but the central role of ChatGPT in formulating the specific step-by-step takeover strategy makes it an AI safety/overreliance incident.

A South Korean gaming company used ChatGPT to formulate a bad-faith strategy to execute a corporate takeover and avoid a 250 million dollar earnout. While highlighting corporate governance risks and overreliance on generative AI for legal strategy, the incident is a private commercial dispute with negligible national security implications.

• Overall national security impact: Negligible
• Scope: Multiple nations
• Primary target: United States
• Other affected: South Korea
• Alleged perpetrator: Krafton Inc.

Threat characteristics

• Imminence: Long-term. The incident is a resolved corporate legal dispute with no immediate or developing national security threat.
• Autonomy: Human-controlled. The AI acted strictly as an assistant to suggest strategies, while human executives retained full decision-making and execution authority.
• Novelty: Evolved capability. Demonstrates an evolved use of consumer LLMs to systematically formulate complex, bad-faith corporate and legal strategies.

Impact by dimension

• Physical security: Negligible. The incident involves a corporate contract dispute in the gaming industry and poses no physical threats, kinetic risks, or critical infrastructure impacts.
• Information security: Negligible. No intelligence compromise, state-sponsored disinformation, or information warfare operations occurred.
• Sovereignty: Negligible. The dispute was resolved standardly within the Delaware Chancery Court, presenting no threat to state authority or sovereign government functions.
• Economic security: Negligible. While involving a 250 million dollar corporate dispute, the incident is limited to the entertainment software sector and does not threaten national economic stability or strategic technology.
• Societal stability: Negligible. The wrongful termination of three executives affects individual employment contracts but does not impact population-scale societal stability or systemic human rights.