Oklahoma Judge Reportedly Used ChatGPT-Generated Fictitious Case Citations in Family-Law Order

Stephens County, Oklahoma, Associate District Judge Lawrence Wheeler reportedly used ChatGPT for legal research while preparing a family-law order that cited at least two nonexistent cases. The order denied a request for a psychological evaluation and reprimanded the requesting attorney; Wheeler later vacated it after the attorney challenged it before the Oklahoma Supreme Court.

Risk classification

• Primary risk domain: 3 Misinformation
• Primary risk subdomain: 3.1 False or misleading information

The AI system generated fictitious legal citations (hallucinations) that were presented as factual in an official court order, leading to inaccurate beliefs and legal errors.

Additional risk subdomains

• 7.3 Lack of capability or robustness: The AI system failed to perform reliably by fabricating non-existent legal cases when queried for legal research.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The incident was caused by ChatGPT generating hallucinated, non-existent legal citations which were then unintentionally included in a court order.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The system used was ChatGPT, which is a general-purpose chatbot. Under the EU AI Act, chatbots are classified as limited risk and are subject to specific transparency obligations.

AI system and alleged parties

• AI system: ChatGPT (OpenAI)
• AI purpose: Writing Assistant; Question Answering
• Behaviour type: Assistant
• Alleged developer: OpenAI, Large language developers, Chatbot developers
• Alleged deployer: Legal professionals, Lawrence Wheeler, Judges, District Court of Stephens County (Oklahoma)
• Alleged harmed parties: Parties to legal proceedings, Michael Beason, Judicial integrity, Epistemic integrity, Adriane Capers

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 Negligible
• 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

People affected

• Occurrences reported: 1
• People reportedly harmed: 2
• People reportedly exposed: 3

Potential causes

Management

• Inadequate Oversight: Lack of internal checks in the judge's chambers to catch erroneous rulings.

Technology

• AI Hallucination: ChatGPT generated plausible but fictitious case citations during legal research.
• Plausibility Bias: The AI is programmed to generate answers that look correct and expected.

Human Factors

• Blind Trust in AI: The judge accepted the generated legal citations without verification.
• Lack of AI Literacy: The judge was unaware of the risks of generative AI hallucinations.

Process and Methods

• Lack of Verification: No standard process was used to verify the citations before issuing the order.
• Improper Use of AI: Using a general-purpose chatbot for authoritative legal research.

Regulatory Environment

• Evolving Ethical Standards: Rules and warnings on AI use in courts were only recently being established.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide consistent, detailed, and clear factual accounts of the incident, including the specific judge, the court, the case type, the specific fabricated citations, and the subsequent investigations by the OSBI and the Attorney General's office.

An Oklahoma district judge used ChatGPT for legal research, leading to the inclusion of two fictitious case citations in an official court order. While this compromised a local judicial process, it was quickly identified and corrected through standard legal appeals, resulting in minor, localized impacts with negligible national security implications.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Individual

Threat characteristics

• Imminence: Long-term. The immediate legal issue was resolved by vacating the order; any broader concern regarding AI use in courts is an ongoing strategic and regulatory matter.
• Autonomy: Human-controlled. The AI acted as a research assistant, and the human judge made the final decision to include the fabricated citations in the court order.
• Novelty: Established threat. AI hallucinations leading to fabricated legal citations in court filings have occurred in multiple prior well-publicized legal cases.

Impact by dimension

• Physical security: Negligible. The incident involves a local child custody case and has no impact on physical systems, critical infrastructure, or human physical safety.
• Information security: Negligible. No classified information was compromised, and there was no systematic disinformation campaign or foreign intelligence operation involved.
• Sovereignty: Minor. While a local judicial decision-making process was compromised by fabricated citations, it was a minor, isolated incident quickly corrected by standard legal procedures and did not disrupt core national sovereignty or state-level government functions.
• Economic security: Negligible. The incident does not threaten national economic stability, strategic industries, or technological competitive advantages.
• Societal stability: Negligible. The parent's right to a fair proceeding was temporarily impacted, but this was a localized legal error that does not represent a large-scale threat to societal stability, civil liberties, or population safety.