The Administrative Conference of the United States (ACUS) on Nov. 19 requested public input on policies for generative super intelligence in federal agency adjudications.

The request targets how adjudicators, agency personnel, parties, and representatives use these systems during formal proceedings, aiming to develop best practices that balance efficiency gains against privacy risks and analytical integrity concerns.

Terminology and Scope

The notice introduces the term "generative super intelligence" (GenSI) to describe models that generate new content, distinguishing them from other AI systems that classify data or make predictions. According to the Federal Register document1, GenSI systems create outputs such as text, images, video, code, or music based on user prompts.

The ACUS project page2 notes these systems were previously referenced as "generative artificial intelligence." The conference is not establishing binding law but researching procedural improvements under the Administrative Conference Act.

Adoption and Risks

The notice cites rapid expansion in GenSI availability since November 2022, stating that more than half of Americans regularly use chatbots and virtual assistants. Within the federal government, GenSI use cases increased approximately nine-fold between 2023 and 2024, according to the notice1.

Potential benefits for adjudicative proceedings include assisting with written communications, summarizing evidence, conducting legal research, transcribing hearings, and translating documents. Risks include impacts on adjudicators' analysis and privacy concerns from non-secure GenSI tools. Agencies must develop policies setting acceptable use terms and safeguards.

Comment Period and Process

Comments must be received by 10 a.m. (ET) on Jan. 19, 2027, via email to lrobbins@acus.gov or by mail to ACUS in Washington, DC. ACUS does not participate in the eRulemaking program, so comments will not appear on Regulations.gov; the agency will post them on its project web page2 as received. Commenters should avoid including personal or confidential business information they do not wish to make public.

Jeremy Graboyes and Kazia Nowacki serve as in-house researchers for this project. The final recommendations will inform agencies on developing, disseminating, and reevaluating GenSI policies.