Michigan Appeals Court Sanctions Attorney for AI Hallucinations

Attorney faces penalties for repeated AI errors in legal filings

A Michigan attorney is facing financial sanctions and a professional misconduct referral after using artificial intelligence to cite non-existent legal precedents in court documents — then compounding the problem by using AI again to correct his mistake.

The Michigan Court of Appeals issued a published opinion on June 17 ordering sanctions against Ronnie Cromer Jr., a Southfield-based attorney, and referring him to the state's Attorney Grievance Commission. Court records show Cromer initially submitted an appeal containing citations to cases that did not exist, a problem that has emerged as a significant risk when lawyers rely on generative AI tools without verification.

The double error

When confronted about the fabricated case citations, Cromer acknowledged he had used artificial intelligence and attempted to file a correction. However, the appeals court determined that Cromer's corrective filing also relied on AI-generated content, resulting in additional erroneous citations.

This pattern of repeated reliance on unverified AI output demonstrated to the court a failure to meet basic professional standards for legal research and citation accuracy. The court's decision to both impose financial penalties and trigger a grievance process reflects the seriousness with which Michigan's judiciary views such lapses.

Why it matters

This case illustrates a growing challenge for the legal profession as AI tools become more accessible. While generative AI can assist with drafting and research, these systems are known to "hallucinate" — confidently producing plausible-sounding but entirely fictional information, including case names, citations, and legal holdings. Attorneys remain professionally and ethically responsible for verifying every citation and fact in their filings, regardless of the tools used to generate initial drafts.

The Michigan court's decision to publish its opinion signals an intent to establish clear precedent: using AI without proper verification is not an acceptable excuse for submitting false information to the court. The referral to the Attorney Grievance Commission could result in additional professional consequences beyond the financial sanctions, potentially including suspension or other disciplinary measures.

For law firms and corporate legal departments adopting AI tools, this case underscores the need for robust verification protocols and training on the limitations of generative AI systems.

These details were first reported by the Detroit Free Press.