AI-related errors in U.S. court filings are no longer isolated incidents but a rapidly expanding trend. The number of documented AI filing mistakes dramatically increased from 25 cases in early 2025 to 249 by the fourth quarter of that year. This pace has continued into 2026, with the first quarter of 2026 nearly matching the entire final quarter of 2025. California leads the nation in these errors, followed by New York, Texas, Florida, and Illinois, accounting for about 40% of all incidents. Florida, in particular, saw a sharp rise in recorded AI-related filing errors during 2025. A significant finding is that over 60% of these documented incidents involved pro se litigants, not attorneys. ChatGPT was the most frequently named AI platform when identified, though most filings did not specify which AI system was used. The most common errors involved outright fabrication, with more than half consisting of nonexistent cases, citations, or legal authorities. Over 80% of these AI-related hallucinations occurred in case-law citations rather than statutes or regulations. Ultimately, AI can be a valuable legal tool, but all citations and legal authorities must be independently verified before court submission.