As per the notice published in the Federal Register on November 19, the Office of the Chairman of the Administrative Conference of the United States (ACUS) is seeking public input on policies for generative AI in federal agency adjudications. The notice uses the term "generative super intelligence", which it says was previously referenced as generative artificial intelligence, and it cites Executive Order 14434 as directing agencies to use that wording. Responses may inform an ongoing ACUS project that will recommend best practices for agencies. The notice covers use by adjudicators and other agency personnel, and by parties and representatives. It lists risks including hallucinated or nonexistent legal authorities cited in filings, and concerns that generative tools could be used to fabricate deepfake evidence.
It says the Executive Office for Immigration Review, the US Patent and Trademark Office and a Department of Labor administrative law judges' standing order of July 8, 2026 already address use by parties and representatives, and that some adjudicators require disclosure or certification. Ten questions ask, among other things, whether agencies should require disclosure of AI use or certification that output was reviewed for accuracy, what consequences improper use should carry, and whether the same rules should apply to lawyers and self-represented parties. The USPTO guidance it cites concerns AI tools in practice before that office, where Indian applicants and counsel also file. Comments are due by 10 a.m. Eastern time on January 19, 2027.
Comments
0 commentsNo comments yet — be the first.