USPTO Issues First Discipline Order Over AI-Generated Hallucinated Citations
The U.S. Patent and Trademark Office (USPTO) has issued its first discipline order predicated on the use of artificial intelligence in legal proceedings. The case centers on "hallucinated" citations—fabricated references generated by AI tools that do not exist in the actual record.
In this instance, a patent practitioner submitted filings containing citations to documents that the AI tool appears to have invented rather than retrieved from the intrinsic record. Unlike general AI errors, citations to the intrinsic record refer specifically to evidence already present in the case file, making the fabrication particularly consequential for legal proceedings.
The order signals that the USPTO will hold practitioners accountable for failing to verify AI-generated content, particularly when that content is presented as factual references to official records. Legal professionals using AI assistance in patent matters are now on notice that unverified outputs can carry professional consequences.
This development comes as courts and agencies across jurisdictions grapple with establishing standards for AI use in legal practice. The USPTO's action provides one of the first concrete examples of regulatory enforcement specifically targeting AI hallucinations in legal submissions, potentially setting a precedent for other administrative bodies and state bar associations.