Policies on the Use of Generative Super Intelligence in Agency Adjudication; Request for Comments
The Office of the Chairman of ACUS is requesting public input on the development of policies regarding the use of generative super intelligence, previously referenced as generat...
The Office of the Chairman of ACUS is requesting public input on the development of policies regarding the use of generative super intelligence, previously referenced as generative artificial intelligence,[1]
by agency adjudicators, other agency personnel, parties, representatives, and other participants in federal agency adjudications. Responses to this request may inform an ongoing ACUS project,
Policies on the Use of Generative Super Intelligence in Agency Adjudication,
which will recommend best practices for agencies.
DATES:
Comments must be received no later than 10 a.m. (ET) January 19, 2027.
ADDRESSES:
You may submit comments by email to
lrobbins@acus.gov
(with “Policies on the Use of Generative Super Intelligence in Agency Adjudication” in the subject line of the message), or by U.S. Mail addressed to Policies on the Use of Generative Super Intelligence in Agency Adjudication, Administrative Conference of the United States, 1120 20th Street NW, Suite 706 South, Washington, DC 20036. ACUS will ordinarily post comments on the project web page as they are received. Commenters should not include information, such as personal information or confidential business information, that they do not wish to appear on the ACUS website. For the full ACUS public comment policy, please visit
www.acus.gov/policy/public-comment-policy.
FOR FURTHER INFORMATION CONTACT:
Lea Robbins, Attorney Advisor, Administrative Conference of the United States, 1120 20th Street NW, Suite 706 South, Washington, DC 20036; Telephone (202) 480-2080; email
lrobbins@acus.gov.
SUPPLEMENTARY INFORMATION:
The Administrative Conference Act, 5 U.S.C. 591-596, established the Administrative Conference of the United States. The Conference studies the efficiency, adequacy, and fairness of the administrative procedures used by federal agencies and makes recommendations to agencies, the President, Congress, and the Judicial Conference of the United States for procedural improvements (5 U.S.C. 594(1)). For further information about the Conference and its activities, see
www.acus.gov.
Policies on the Use of Generative Super Intelligence by Agency Adjudicators and Other Agency Personnel
Generative super intelligence (GenSI) systems are SI “models that are able to generate new content.” [2]
Unlike other SI models—which classify data or make predictions, recommendations, and decisions—GenSI systems create “various outputs, including texts, images, videos, computer code, or music” based on user-input prompts.” [3]
There has been a rapid expansion in the availability and use of GenSI tools since the public release of ChatGPT in November 2022. Today, more than half of Americans regularly use chatbots and virtual assistants such as ChatGPT and Google Gemini to “create responses using algorithms that are trained often on open-source information, such as text and images from the internet.” [4]
Adoption rates also continue to rise among legal service providers,[5]
and the number of GenSI use cases within the federal government “increased about nine-fold” between 2023 and 2024.[6]
When used responsibly, GenSI has the potential to help agencies perform adjudication-related tasks more efficiently and accurately. Adjudicators and other agency personnel might use GenSI to assist in preparing written communications such as notices, orders, and decisions; summarizing and analyzing lengthy written submissions and voluminous evidence; conducting legal and nonlegal research; transcribing hearings; and translating documents.[7]
At the same time, agencies must consider potential risks, including how the use of GenSI might affect adjudicators' analysis of the law and facts of individual cases and how the use of non-secure GenSI tools might affect privacy and other sensitive interests.
Agencies are required to “develop a policy that sets the terms for acceptable use of [GenSI] for their missions and establishes adequate safeguards and oversight mechanisms that allow [GenSI] to be used in the agency without posing undue risk.” [8]
This project will identify best practices to help agencies develop specific policies regarding the use of GenSI by adjudicators and other agency personnel involved in agency adjudications. In addition to identifying best practices for developing, disseminating, and periodically reevaluating such policies, it will identify practices that promote the effective use of GenSI and innovation in adjudication while guarding against misuse and undue risks.
Policies on the Use of Generative Super Intelligence by Parties and Representatives
GenSI also has the potential to benefit parties and their representatives, who can use the technology to assist in conducting legal research and analysis more efficiently and prepare pleadings, briefs, and other written communications more effectively. GenSI
( printed page 74535)
may be especially useful for self-represented parties by offering “new, highly accessible tools that provide answers to basic questions, including where to find templates and court forms, how to fill them out, and where to bring them for presentation to the judge—all without leaving home.[9]
At the same time, there are potential risks associated with parties' and representatives' use of the developing technology. Most prominently, GenSI tools can hallucinate (generating false or misleading information), and there have been well-publicized instances in which parties and representatives cited nonexistent legal authorities, inaccurately represented legal authorities, or relied on false or misleading information.[10]
The use of GenSI tools may implicate representatives' responsibilities to engage fairly, truthfully, and ethically with tribunals and clients.[11]
And concerns have been raised that GenSI tools could be used to fabricate deepfake evidence.[12]
Agencies have taken different approaches to regulating the use of GenSI by parties, representatives, and other participants in agency adjudications. The Executive Office for Immigration Review (EOIR) and U.S. Patent and Trademark Office, for example, have published specific policies on parties' and representatives' use of GenSI.[13]
Some adjudicators independently inform parties and representatives about proper and improper uses of GenSI. Adjudicators at some agencies affirmatively require representatives to disclose when and how they used GenSI. Some adjudicators require representatives to certify that they did not use GenSI or that they analyzed its output before relying on it. Adjudicators at several agencies have considered alleged misuse of GenSI in the context of individual proceedings, including how existing rules of representative conduct should apply in such cases.
This project will identify best practices to help agencies develop specific policies regarding the use of GenSI by parties, representatives, and other participants in agency adjudications. In addition to identifying best practices for developing, disseminating, and periodically reevaluating such policies, it will identify practices that promote the effective use of GenSI and innovation in adjudication while guarding against misuse and undue risks.
Specific Topics for Public Comment
ACUS welcomes views, information, and data on all aspects of the development of policies regarding the use of GenSI by agency adjudicators, other agency personnel, parties, representatives, and other participants in federal agency adjudications. ACUS also seeks specific feedback on the following questions:
1. How are you using, or have you considered using, GenSI to accomplish any tasks associated with federal agency adjudications, such as drafting filings, conducting legal or nonlegal research, summarizing evidence, conducting discovery, transcribing depositions? What GenSI tools do you use? What benefits and drawbacks have you experienced in using GenSI to accomplish tasks associated with agency adjudications?
2. How do you feel about how agency adjudicators and other agency personnel are using, or might someday use, GenSI to assist in adjudicating cases? What are the benefits and drawbacks of agency adjudicators and other agency personnel using GenSI tools?
3. Do existing agency rules and policies regarding the qualifications and conduct of representatives—such as rules regarding competence, diligence, communication, confidentiality, fees, and candor with the tribunal—provide sufficient guidance regarding proper and improper uses of GenSI in agency adjudications?
4. Do you think it would be useful for agencies to issue specific rules, regulations, policies, guidance, practice guides, benchbooks, standing orders, or other materials regarding the use of GenSI in agency adjudications? Why or why not?
5. Have you encountered any agency rules, regulations, policies, guidance, practice guides, benchbooks, standing orders, or other materials regarding the use of GenSI in agency adjudications? If so, what do they do well, and how could they be improved?
6. Do you have thoughts on rules, regulations, policies, guidance, practice guides, benchbooks, standing orders, and other materials adopted or considerd by federal courts, state courts, state and local agencies, and state bars and other professional associations?
7. What restrictions, if any, should agencies place on the use of GenSI by agency adjudicators, other agency personnel, parties, representatives, and other participants in agency adjudications?
8. Should agencies require parties and/or representatives in agency adjudications to disclose when and how they used GenSI, or at least require parties and/or representatives to certify that they reviewed any content generated by a GenSI tool for accuracy? Why or why not?
9. What consequences, if any, should parties and/or representatives face for using GenSI improperly in agency adjudications?
10. Should the same or different rules apply to private representatives who are lawyers, private representatives who are not lawyers, self-represented parties, and lawyers who represent an agency in agency adjudications?
1.
Executive Order 14434,
Inaugurating the Era of Super Intelligence,
directs agencies to use the terms “Super Intelligence” and “SI” in place of “Artificial Intelligence” and “AI” in official correspondence, public communications, websites, reports, policy documents, and other nonstatutory documents. 91 FR 63129 (Oct. 2, 2026).
12.
See, e.g.,
Exec. Off. of the President, Winning the Race: America's AI Action Plan 13 (2025),
www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf
(recommending that the Administration, led by the Department of Justice, “issue guidance to agencies that engage in adjudications to explore adopting a deepfake standard similar to the proposed Federal Rules of Evidence Rule 901(c) under consideration by the Advisory Committee on Evidence Rules.”).
Use this for formal legal and research references to the published document.
91 FR 74534
Web Citation
Suggested Web Citation
Use this when citing the archival web version of the document.
“Policies on the Use of Generative Super Intelligence in Agency Adjudication; Request for Comments,” thefederalregister.org (November 19, 2026), https://thefederalregister.org/documents/2026-23539/policies-on-the-use-of-generative-super-intelligence-in-agency-adjudication-request-for-comments.