A director and the president of one of Delaware's oldest law firms apologized to the state's Chancery Court on Wednesday for submitting a legal brief that contained fabrications produced by a generative AI tool. The firm, Richards, Layton & Finger PA, asked the court not to impose sanctions, acknowledging the Jan. 22 filing and a subsequent correction that failed to fully remove the errors burdened the court.
Richard Rollo, a director at the firm, and Paul Heath, the firm's president, each filed affidavits expressing regret. "I regret the negative impression that I have made," Rollo said. "While I used a GenAI tool, I had in place a human-centric process (starting with me) that should …" The quoted statement cuts off in the court filing, but Rollo's admission makes clear the firm relied on AI-generated content without adequate verification.
A cascade of errors in Delaware's Chancery Court
The incident traces back to a brief filed on Jan. 22 that contained hallucinated case citations and legal arguments. A follow-up filing meant to correct the record did not completely scrub the fabrications, compounding the problem. Heath told the court the firm had "burdened the court" with the filings and accepted responsibility for the lapses in the review process.
Richards, Layton & Finger is one of the largest and most established law firms in Delaware, a jurisdiction that handles massive volumes of corporate litigation. The court's response to the request for no sanctions will be closely watched by legal professionals who increasingly face pressure to adopt AI tools while safeguarding accuracy.
AI hallucinations put legal workflows under scrutiny
This case is not the first time generative AI has produced false legal content, but the prominence of the firm and the court amplifies the warning. AI systems can fabricate cases, statutes, and quotes with confident phrasing, making errors hard to spot without rigorous human review. The incident underscores the gap between the speed of AI adoption and the slow, deliberate work of authenticating legal authority.
For professionals seeking to understand the risks and proper use of these tools, Generative AI and LLM Training covers the mechanics behind such hallucinations and how to build verification into workflows. The core lesson from this Delaware filing is that a "human-centric process" must mean more than a quick review-it requires systematic checks against primary sources.
Why this matters for legal professionals
Courts are losing patience with AI-generated errors. Sanctions, reputational damage, and malpractice claims are real possibilities when hallucinations slip through. The Delaware case shows that even large firms with substantial resources can fail to catch these mistakes. Building reliable AI use protocols-including mandatory verification steps and training on the limitations of generative tools-is no longer optional. For tailored guidance, AI for Legal Professionals provides practical strategies for integrating AI into legal research and drafting while maintaining ethical and professional standards.
Your membership also unlocks:
