Skip to content
Submit incident
AII-1000-67843
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.Supporting sources are pending verification.

Sullivan & Cromwell court filing contained AI-generated legal errors

Sullivan & Cromwell apologized to a federal bankruptcy judge after filing a motion on 9 April containing around three dozen errors generated by artificial intelligence, including fabricated case citations and misquoted legal authorities. The errors were discovered by opposing counsel, and the firm acknowledged that its AI policies were not followed and its secondary review process failed to catch the mistakes.

9 April 2026 Β· Sullivan & Cromwell Β· Reliability and hallucination
9 April 2026
AII-1000-67843
Quick facts
Category
Reliability and hallucination
Country
United States
Date
9 April 2026
Drafted by
Raidu Incident Detection

What happened

Sullivan & Cromwell filed a motion in U.S. Bankruptcy Court in Manhattan on 9 April containing errors generated by artificial intelligence, including inaccurate citations and misquoted passages from the U.S. bankruptcy code.

The errors, totaling around three dozen, were discovered by lawyers from Boies Schiller Flexner. Andrew Dietderich, co-head of S&C's global restructuring group, apologized to Judge Martin Glenn in a letter on 18 April and said the firm's AI policies were not followed and its secondary review process failed to catch the errors.

Reported impact

The filing errors affected a case involving S&C's representation of liquidators in actions against Prince Group, owned by Chen Zhi, who was charged by US prosecutors with wire fraud and money laundering.

Index analysis

Sullivan & Cromwell states it maintains comprehensive policies and training requirements governing AI tool use in legal work, but acknowledged those policies were not followed in preparing this motion and did not prevent the hallucinated citations from reaching the court.

The firm later filed a corrected version to the court. The letter did not disclose which AI program was used or which lawyers prepared the document.

Governance controls

The American Bar Association has instructed lawyers to exercise caution when using AI models. Lawyers are ethically bound to ensure the accuracy of court submissions.

Classification

Category
Reliability and hallucination
Country
United States
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Public interest, Reputational
Failure modes
Fabricated citation, Hallucination
Consequences
Public apology
AI systems
Unidentified AI tool

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident