The scope of the problem remains incompletely documented. While hundreds of reported instances exist worldwide with increasing frequency in 2025, a comprehensive accounting of all AI-related citation errors in active litigation does not appear to be publicly available. The specific frequency with which courts are encountering these errors, and whether sanctions are being imposed consistently across jurisdictions, is still developing.
Attorneys should treat AI-generated legal research as requiring the same citation verification they would apply to any unfamiliar source. Courts are now imposing sanctions for AI-generated errors, and bar referrals and mandatory training requirements are becoming more common. The risk extends beyond malpractice exposure: filing false citations undermines judicial confidence in the profession itself. Any firm deploying AI for research or drafting should implement mandatory citation-checking protocols and ensure those policies are actually followed.