概要: Ko v. Li, 2025 ONSC 2766において、オンタリオ州高等裁判所は、申請者側の弁護士が提出した事実陳述書(factum)において、複数の不正確な、あるいは存在しない判例引用を特定しました。F.L.マイヤーズ判事は、当該文書の作成に生成AIが使用されたかどうかを疑問視し、引用された判例は無関係、不正確、あるいは存在しないと指摘しました。裁判所は、正確性義務と技術的能力を理由に、弁護士に対し、侮辱罪で告発されるべきでない理由を示すよう命じました。
Editor Notes: Timeline note: April 25, 2025 is the date the factum containing allegedly AI-generated citations was signed and submitted. May 1 marks the court hearing where the judge identified and questioned the suspect citations. May 6 is the date of the written decision ordering counsel to show cause. Clarification regarding the AI system and deployer: Justice Myers writes, "I asked Ms. Lee if her factum was prepared by artificial intelligence - like ChatGPT. She told me that her office does not usually do so but that she would have to check with her clerk. Ms. Lee was unable to provide me with citations to the cases cited in her factum or to provide me with copies of the cases from the printed papers she was using to make her submissions." The incident is tagged with "Unspecified large language model" as the AI system, since "like ChatGPT" is used illustratively, not as a definitive attribution. While Jisuh Lee is tagged as the deployer, her recorded answer suggests the possibility that someone else may have deployed the large language model to generate the citations.
Alleged: Unspecified large language mode developer developed an AI system deployed by Jisuh Lee, which harmed Jisuh Lee , Hanna Ko , Ontario Superior Court of Justice と Legal integrity.
関与が疑われるAIシステム: Unspecified large language model
インシデン トのステータス
インシデントID
1099
レポート数
2
インシデント発生日
2025-04-25
エディタ
Daniel Atherton