Sullivan & Cromwell Submits AI-Generated Errors in Bankruptcy Filing

This incident highlights the systemic risks associated with AI adoption in high-stakes legal environments.

• On April 9, 2026, Sullivan & Cromwell filed an emergency motion in U.S. Bankruptcy Court that contained approximately 36 errors, including AI-generated hallucinations.
• By April 18, 2026, partner Andrew G. Dietderich issued an apology to Judge Martin Glenn, admitting to policy violations and submitting a corrected version of the motion.
• As of April 22, 2026, no sanctions have been imposed, but the incident has drawn significant media attention and raised questions about AI's reliability in legal contexts.

• Prince Global Holdings Limited, involved in a bankruptcy case, is linked to allegations of scam operations and nearly $9 billion in frozen bitcoin.
• Sullivan & Cromwell, a prestigious Wall Street law firm, has faced scrutiny for its use of AI in legal filings, reflecting broader challenges in the legal industry since 2023.
• Boies Schiller Flexner LLP, the opposing counsel, discovered the errors and notified Sullivan & Cromwell, emphasizing the importance of verification in legal documents.

The incident involving Sullivan & Cromwell serves as a cautionary tale about the integration of artificial intelligence in legal practices. As AI tools become more prevalent in drafting legal documents, the risk of errors—especially those stemming from AI hallucinations—has emerged as a significant concern. Hallucinations refer to instances where AI generates information that appears plausible but is entirely fabricated. In this case, the emergency motion filed by Sullivan & Cromwell contained fabricated case citations and misquotes of U.S. bankruptcy law, which were identified by opposing counsel.

The reliance on AI in legal contexts is driven by the need for efficiency and cost reduction. Law firms are under pressure to deliver faster results while managing increasing workloads. However, this incident illustrates the potential pitfalls of prioritizing speed over accuracy. The errors in the filing not only jeopardized the integrity of the case but also raised questions about the firm's adherence to internal policies regarding AI usage. Andrew Dietderich's apology to Judge Glenn highlighted a failure in the review process, suggesting that the firm did not adequately vet the AI-generated content before submission.

This incident is not isolated; it reflects a broader trend in the legal industry where firms are increasingly adopting AI tools without fully understanding the implications. Since 2023, there have been multiple instances where legal professionals faced fines for submitting documents with fabricated citations generated by AI. The Sullivan & Cromwell case underscores the urgent need for robust verification processes and human oversight in AI applications.

Moreover, the lack of immediate sanctions following the incident may indicate a reluctance within the legal system to impose strict penalties on elite firms, which could further entrench the use of AI without adequate safeguards. As firms continue to navigate the complexities of AI integration, the balance between leveraging technology for efficiency and maintaining the integrity of legal processes will be critical.

• Legal professionals: Increased scrutiny on AI usage may lead to stricter regulations and oversight.
• Law firms: Potential reputational damage and loss of client trust if AI errors occur.
• Clients: Risk of compromised legal outcomes due to reliance on flawed AI-generated documents.

• Increased regulatory scrutiny: Watch for potential new guidelines or regulations governing AI use in legal filings, which could reshape industry standards.
• Adoption of verification technologies: Firms may invest in advanced verification tools to mitigate risks associated with AI-generated content, impacting technology vendors.
• Public perception shifts: Monitor how this incident influences client trust in legal firms that utilize AI, potentially affecting client retention and acquisition strategies.