Bunce v. Visual Technology Innovations, Inc.

What sanction did the court impose?

Two-part 2025 sanction under Rule 11(c): a $2,500 monetary penalty payable by Rajan personally, and a mandatory one-hour CLE-credited seminar on AI and legal ethics. The 2026 follow-on sanction added a $5,000 monetary penalty plus an order requiring Rajan to complete additional CLE on artificial intelligence and legal ethics and to provide proof of any related CLE already completed. Rajan's 2026 request that the court limit the penalty to $950 in light of more than $73,000 in prior sanctions in the case was rejected. Total AI-citation sanctions across the two orders in the same litigation: $7,500.

Why does Bunce v. Visual Technology Innovations, Inc. matter for law firms using AI?

Bunce v. VTI is the clearest case in the dataset where the same attorney is sanctioned twice in the same litigation for AI-fabricated citations. The 2025 sanction is a useful teaching case because the conduct was not strategic misconduct but credulous adoption of an unfamiliar tool: counsel admitted he was new to ChatGPT, had been told about it by a friend, and treated it as interchangeable with the vetted research platform he normally used. The 2026 follow-on, captured in Judge Scott’s line “it will be your name and license on the line, not ChatGPT’s,” signals that courts will escalate, not excuse, repeat hallucination conduct, and that “I relied on the tool” is not a defense the second time. Managing partners should treat a single hallucination incident as a supervision trigger, not a one-off, because the bench is now watching for pattern conduct, and Judge Scott’s framing of the signing attorney as the “final auditor” of every legal claim gives firm administrators a clean policy rationale for requiring documented citation verification regardless of which AI tool a lawyer uses.