Courts across the United States are imposing sanctions on attorneys who submit legal filings containing citations to cases that do not exist — fabrications generated by AI language models like ChatGPT and Claude. The sanctions have been severe: fines reaching $10,000 or more, orders to show cause, dismissal of cases, referrals to state bar disciplinary authorities, and in some cases suspension proceedings. The Ethics Reporter tracks every major AI sanctions case, from the first high-profile incident in the Mata v. Avianca case to the growing wave of sanctions in 2025 and 2026. We examine the cases in detail — who the attorneys were, what AI tools they used, what fabrications appeared in their filings, and what sanctions resulted. But we also ask the harder question: are these sanctions proportionate, or are courts using AI misconduct as an opportunity to make an example of attorneys in ways that serve institutional interests rather than justice?