Spanish Court Fines Lawyer Over 48 AI-Fabricated Rulings in Appeal

The Criminal Chamber of the Superior Court of Justice of the Canary Islands (Tribunal Superior de Justicia de Canarias, TSJC) fined a lawyer 420 euros after he cited 48 rulings in a cassation appeal that had, in fact, never been issued. The rulings' texts, case numbers, dates and identifiers were invented by a general-purpose AI chatbot, and the lawyer filed them with the court without any verification.

The case concerned a cassation appeal filed with the TSJC's Criminal Chamber against a ruling previously issued by the Court of Appeal in Santa Cruz de Tenerife. In the filing, the lawyer cited several dozen precedents meant to support his legal argument. When the court checked the cited rulings, it found that the vast majority simply did not exist.

A general tool, not a legal one

The lawyer had not used specialized AI-based legal software, but a general conversational tool not designed to work with case law databases. Such systems are known to hallucinate, generating plausible-sounding but fictitious content when they lack access to actual source data.

The court stressed in its reasoning that the lawyer never cross-checked the ruling numbers, dates or case identifiers against any official case law database before filing the document. Instead, he transferred the chatbot-generated content directly into the court filing.

The duty of human oversight

The lack of verification constitutes a breach of the basic duty of human oversight, which is unavoidable when using artificial intelligence tools in professional practice - Tribunal Superior de Justicia de Canarias

The court noted that using AI tools is not objectionable in itself, but shifting responsibility for the accuracy of cited case law onto the algorithm is unacceptable in a legal representative's professional practice. The judges emphasized that it is the lawyer, not the tool, who bears responsibility for the content of filings submitted to the court.

How the fine was calculated

In setting the fine, the court referred to the cost of an annual subscription to a specialized legal AI tool, estimated at around 840 euros, and imposed half that amount. The judges took into account that such a specialized system, unlike a general chatbot, would likely have prevented the lawyer's error, since it works with verified case law databases. The penalty also took into account that the lawyer admitted his mistake and expressed remorse before the court.

The TSJC also decided to forward the ruling, along with the case file, to the relevant bar association, which will assess whether the lawyer's conduct warrants separate disciplinary proceedings. It is a sign that Spanish courts are increasingly taking such cases seriously, not stopping at a financial fine alone.

A growing problem worldwide

The Canary Islands case is not an isolated one. Researchers tracking such incidents worldwide have already recorded more than thirteen hundred court cases in which fictitious rulings, invented citations or nonexistent AI-generated case law made it into court files, with hundreds of lawyers sanctioned for it across different continents. Courts in the United States, the United Kingdom and other European countries have repeatedly imposed financial penalties on lawyers in similar situations.

For Polish lawyers and law firms, the case is a significant warning, especially given the growing popularity of AI tools in drafting court filings. The Naczelna Rada Adwokacka (Poland's National Bar Council) recently introduced its own rules on AI use in its Code of Ethics, and cases like the Spanish one show that failing to verify AI-generated content can cost not just a fine, but disciplinary proceedings as well.

In its reasoning, the court clearly distinguished between using the technology itself and neglecting basic professional diligence. The ruling is expected to feed into a broader discussion about the limits of professional liability for lawyers using generative artificial intelligence in their daily work.