Romanian Lawyer Barred from ECHR for Two Years Over AI-Generated Filings

A Romanian lawyer has been barred for two years from representing clients before the European Court of Human Rights after filing written arguments generated by artificial intelligence that cited phantom cases and fabricated legal passages.

The court’s president imposed the sanction on a proposal from the chamber hearing the case, where the lawyer served as one of two procedural representatives, DeFakto.bg reported.

Submissions drafted by the lawyer contained numerous errors regarding the court’s established practice, according to the court’s announcement. The filings referenced non-existent hallucinated cases, cited made-up paragraphs in actual rulings, and invoked irrelevant precedents that contradicted the arguments they were meant to support.

The chamber rejected that characterization, finding instead a systematic pattern of repeated substantive errors rather than isolated lapses.

Photo: DeFakto.bg

Judges concluded the conduct failed to meet professional and ethical obligations requiring lawyers to act as reasonably competent professionals with a high degree of care. The chamber noted the lawyer failed to independently verify AI-generated references against authoritative sources such as the court’s own database, adding that many errors would have been obvious even during a cursory check.

The misconduct damaged the administration of justice by forcing the court to spend significant time reviewing filings, opening proceedings, and managing extensive correspondence with other parties. The chamber evaluated responses from the lawyer, the Romanian government, and the National Union of Romanian Bar Associations before issuing the prohibition.

“The colleague not only wrote her submission with AI, but did not read at all what the language model returned to her,” Kachanova said, according to DeFakto.bg.

At a recent meeting in Strasbourg between procedural representatives, judges, and court administration, an audience member asked whether the judges themselves use AI in drafting decisions or in the court’s work. The head of the court’s IT department replied that the court’s bigger problem is the growing number of lawyers using AI in applications and submissions, and that if judges uphold work standards, that does not apply to procedural representatives.