DE | Amtsgericht Köln
Docket: 312 F 130/25
ECLI:DE:AGK:2025:0702.312F130.25.00 • first instance
2025-07-02
amtsgericht köln — beschluss vom 2. juli 2025 — 312 f 130/25
Unknown / Not Disclosed (artificial intelligence mentioned) ai generated legal argumentfake case citationwrong citation to real sourceno human verification
I. Executive Summary
In a family-law order, the court stated that certain 'prerequisites' attributed to a cited high-court decision were not contained in that decision and were apparently generated by artificial intelligence and invented.
The court also stated that the cited sources and pinpoint references were invented and corrected multiple misattributions.
The court warned counsel to refrain from such submissions because they hinder legal reasoning and mislead readers.
It also referenced the professional duty not to spread untruths in submissions.
II. Conduct Analysis
Counsel included in a written submission AI-generated and invented legal prerequisites and citations (including invented sources and pinpoint references) attributed to a high-court decision.
The court identified the inaccuracies and admonished counsel not to repeat such conduct.
III. Legal Foundations
Section 43a(3), federal lawyers’ act (BRAO)
IV. Key Facts
1) The court stated that additional prerequisites cited in a submission were not from the referenced decision and were apparently generated by artificial intelligence and freely invented, with invented citations.
2) The court corrected multiple misattributions (commentary authorship, edition years, missing paragraph numbers) and noted several cited works could not be found.
3) The court warned counsel to refrain in the future and pointed to possible violation of section 43a(3) of the federal lawyers’ act (BRAO) regarding spreading untruths.
V. Consequences & Sanction
1) No separate sanction mechanism was ordered in relation to the AI-related conduct within this order, the court issued an explicit warning to counsel.
2) The decision proceeded on the merits of the family-law issues independently of the AI-related commentary.