Holding / Framework
Standing order pioneers the **confidentiality / business-proprietary disclosure-risk framework** for judicial AI governance. Distinct from Starr (accuracy/Rule 11 framework) and Baylson (broad-scope verification framework). Vaden's framework recognizes a second category of AI-use risk beyond hallucination: that prompt inputs may transmit otherwise-protected litigation information to third-party AI providers.
Triggering Conduct
Issued in response to growing concern over generative-AI tools producing fabricated citations AND, distinctly from other early standing orders, the inadvertent disclosure of business-proprietary information through AI prompt inputs (e.g., where confidential trade data is included in prompts and effectively transmitted to third-party AI providers).
Disclosure Requirement
Two-prong requirement for any submission containing text drafted with the assistance of a generative AI program based on natural-language prompts (including but not limited to ChatGPT and Google Bard): (1) DISCLOSURE NOTICE, must identify the program used and the specific portions of text that were AI-drafted; (2) CONFIDENTIALITY CERTIFICATION, must certify that use of the program has not resulted in disclosure of any confidential or business-proprietary information to any unauthorized party.
Primary Source
https://www.cit.uscourts.gov/sites/cit/files/Order%20on%20Artificial%20Intelligence.pdf · court_website
Tags
specialty_court, court_of_international_trade, standing_order, disclosure_requirement, certification, confidentiality_framework, business_proprietary_risk, framework_setter