Expert: Damien Charlotin, Legal Researcher, Artificial Authority AI-hallucinated legal filings have become a global epidemic in courtrooms. Damien Charlotin, a legal researcher who maintains the most comprehensive database of AI hallucination cases in law, has documented **1,782 cases** across courts in 39 countries as of July 2026. The database reveals that AI models — including ChatGPT, Gemini, and others — routinely fabricate case citations, invent judicial quotations, and generate non-existent legal sources that are then submitted to real courts. Both lawyers (695 cases) and pro se litigants (1,043 cases) have been caught. **Scale by country:** USA leads with 1,228 cases, followed by Canada (198), Australia (96), UK (60), Israel (54), and India (12). **Nature of hallucinations:** Fabricated case law (1,599 cases), misrepresented legal norms (737), false quotations (484), and fabricated exhibits (126). **Consequences:** Courts worldwide are imposing increasingly severe sanctions — monetary penalties up to $29,877 in a single case, bar referrals, practice bans, and orders to complete CLE courses on AI ethics. **Landmark recent cases:** - **Supreme Court of India** (July 2026) set aside multiple orders after discovering the National Company Law Tribunal had relied on AI-generated non-existent case law, declaring it "a subversion of the rule of law." - **Spain** — an innovative court fined a lawyer €1,800, calibrated to the annual cost of a verified legal AI tool, reasoning that using proper AI tools would have prevented the hallucination. - **USA** — In Re Rosslyn2016 (Texas, July 14, 2026): a lawyer using Westlaw Precision submitted fabricated case citations and was ordered to pay $29,877 in costs and complete CLE on generative AI. The database, featured in the LA Times, Volokh Conspiracy, Reason, and 404 Media, is updated daily and now includes 47 new cases in just the first 18 days of July 2026 alone — showing the problem is accelerating.