Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Ai Governance topic
No spam. Unsubscribe anytime.
Judicial Council adopts rule and standard governing generative AI in court work; disclosure, privacy, and verification required
Summary
The Judicial Council adopted a new rule and accompanying standard to govern the use of generative artificial intelligence in court‑related work, requiring court policies on confidentiality, verification, bias mitigation, and disclosure for public AI‑generated materials.
The Judicial Council adopted a rule of court (proposed rule 10.43) and a standard of judicial administration to address the risks of generative artificial intelligence (AI) in court‑related work. The action sets baseline safeguards while allowing courts discretion to permit, restrict, or prohibit AI use locally.
Administrative Presiding Justice Hill, chair of the Chief Justice’s Artificial Intelligence Task Force, presented the recommended rule and standard. The rule requires courts that do not prohibit generative AI for court work to adopt a written use policy by Dec. 15, 2025. Required policy elements include protections for confidentiality and privacy; supervision, verification, and accountability to address AI “hallucinations” (fabricated or inaccurate outputs); measures to prevent or remove biased, offensive, or discriminatory content; and a disclosure requirement for public materials wholly generated by AI. Justice Hill explained the rule’s transparency requirement: when written, visual, or audio work that will be provided to the public consists entirely of generative AI output, courts must label or otherwise disclose how AI was used and identify the system employed.
The rule bars submitting confidential personal identifying information or other non‑public data to generative AI systems not controlled or managed by the court and requires users to take reasonable steps to verify and correct AI output. The companion standard guides judicial officers using generative AI for adjudicative tasks, advising they avoid entering confidential identifying data, verify AI output, avoid unlawful discrimination, and consider disclosure to the public when AI produces content provided to the public.
The task force said the rule and standard focus on risk management rather than an exhaustive list of permitted tools; it will publish a model policy and continue monitoring technical, ethical, and legal developments, including education and training for officers and staff. The council approved the rule and standard after a motion; the vote was recorded as overwhelmingly in favor with one member, Mr. Pritt, recorded as voting no.
