Sixth Circuit Sanctioned Lawyers in Whiting v. City of Athens over Alleged Fake Appellate Citations in Briefs Reportedly Bearing Hallmarks of Hallucinations

The U.S. Court of Appeals for the Sixth Circuit sanctioned attorneys Van Irion and Russ Egli after reportedly finding more than two dozen fake citations and alleged factual misrepresentations in appellate briefs in Whiting v. City of Athens. The court asked whether generative AI had been used, but the attorneys reportedly did not answer that question. Public reporting described the filings as bearing hallmarks of AI hallucinations.

Risk classification

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

The AI system generated discrete false information in the form of over 24 fabricated legal citations, which were then submitted to a federal court.

Additional risk subdomains

• 5.1 Overreliance and unsafe use: The attorneys overrelied on the AI system's outputs to draft their legal briefs without performing necessary verification or cite-checking.

Causal factors

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

The incident was initiated by an AI system generating unintended hallucinations (fake citations) during post-deployment use by the attorneys.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The system described is a generative AI platform used for text generation and drafting, which falls under Limited Risk (Level 3) as it involves AI-generated content that requires transparency and human verification.

AI system and alleged parties

• AI system: generative AI
• AI purpose: Writing Assistant; Technical Text Generation
• Behaviour type: Assistant
• Alleged developer: Unknown large language model developers, Unknown generative AI developers
• Alleged deployer: Van Irion, Russ Egli
• Alleged harmed parties: U.S. Court of Appeals for the Sixth Circuit, Judicial integrity, Epistemic integrity, City of Athens, Tennessee

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. 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 Minor, 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 Minor, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: The report explicitly describes financial losses in the form of court-ordered sanctions and legal fees.

Directly caused: The two attorneys were directly ordered to pay 15,000 dollars each in punitive sanctions, totaling 30,000 dollars, plus double costs and reimbursement of legal fees.

Inferred additional harm: The total financial impact including double costs and reimbursed attorneys' fees is likely to exceed 50,000 dollars, though the exact total is not fully quantified.

Epistemic

Reported: The report explicitly describes epistemic harm in the form of fabricated legal citations and factual misrepresentations.

Directly caused: The AI system generated, and the attorneys submitted, over 24 entirely fabricated legal citations and false propositions of law to a federal court.

People affected

• Occurrences reported: 1
• People reportedly harmed: 2
• People reportedly exposed: 5

Potential causes

Management

• Inadequate Supervision: Law firm partners failed to supervise the integration of AI tools.

Technology

• Generative AI Hallucinations: AI platforms fabricated fake case citations and misrepresented the law.

Human Factors

• Failure to Vet AI Output: Attorneys failed to check the accuracy of the AI-generated citations.
• Defiance of Court Inquiry: Attorneys refused to answer the court's questions about AI use.
• Lack of Candor before Court: Attorneys had prior discipline for lack of candor before a tribunal.

Process and Methods

• Flawed Cite-Checking Process: Attorneys lacked a robust process to verify the authenticity of citations.
• Deficient Brief Preparation: Attorneys submitted briefs with over two dozen fake citations.

Regulatory Environment

• No Local AI Disclosure Rules: The circuit lacked local rules requiring proactive disclosure of AI use.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide precise details of the court case, the specific sanctions (15,000 dollars each, double costs), and the nature of the fake citations. While the attorneys did not explicitly admit to using AI, the court and reports note the filings bore the clear hallmarks of AI hallucinations, and the legal analysis of the incident is highly documented.
• Ambiguities identified: The attorneys did not explicitly confirm whether they used generative AI, though the court noted the filings bore the hallmarks of AI hallucinations.
• Alternative interpretations: The attorneys might have fabricated the citations manually or through some other non-AI means, though this is highly unlikely given the characteristic nature of the hallucinations.

Two attorneys were sanctioned by a federal court for submitting briefs containing AI-generated fake citations. While this represents a minor disruption to judicial processes (sovereignty and government functions) that was resolved through standard court procedures, it carries negligible national security implications.

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

Threat characteristics

• Imminence: Long-term. The incident is a resolved legal matter with no ongoing or immediate national security crisis.
• Autonomy: Human-controlled. The generative AI system acted as a writing assistant, and humans retained final control over the submission of the briefs.
• Novelty: Established threat. The use of generative AI leading to hallucinated legal citations in court filings is an established threat with several prior documented cases.

Impact by dimension

• Physical security: Negligible. No physical threat, kinetic attack, or critical infrastructure disruption occurred in connection with this legal incident.
• Information security: Negligible. No foreign disinformation campaign, intelligence compromise, or classified data theft was associated with this incident.
• Sovereignty: Minor. The submission of fabricated citations to a federal court disrupted judicial proceedings, representing a minor impact managed through standard court sanction procedures.
• Economic security: Negligible. The incident involved localized financial sanctions on individual lawyers and has no broader implications for strategic industries or national economic security.
• Societal stability: Negligible. This incident did not involve mass surveillance, systematic discrimination, civil unrest, or large-scale threats to societal stability.