The Judiciary’s AI Governance Framework

The framework

The Supreme Court adopted the Governance Framework on the Use of Human-Centered Augmented Intelligence in the Judiciary through its February 18, 2026 Resolution in A.M. No. 25-11-28-SC. Its official release followed on March 19, 2026.

The framework sets standards for introducing and using AI in judicial work. It combines ethical principles with controls for adoption, oversight, risk management, and data governance. It is an administrative issuance rather than a judgment resolving a dispute between litigants.

Who it covers

Part I.B covers justices and judges, court officials and employees, court users, and vendors or organizations developing or using AI for the Judiciary. Court users include lawyers and other members of the public who seek relief from or transact with courts or judicial offices.

Under Part I.C.6(b), the relevant activities include preparing and submitting court-bound documents and presenting evidence. A lawyer using AI to help prepare a pleading therefore needs to consider the framework even when the work takes place outside a courthouse.

Key safeguards

• Authorization. Part II.B.3 states that no AI tool shall be used unless authorized by the Supreme Court En Banc. The framework’s coverage and any applicable implementing directions must be checked before adopting a tool for judicial work. A commercial subscription does not establish that authorization.
• Human review and responsibility. Part II.A.4 requires people to review and approve AI outputs. AI must never provide the sole, primary, or determinative basis of an adjudicatory outcome. Judges must form their own legal reasoning and conclusions; users remain responsible for the work and its consequences.
• Disclosure. Part II.A.3 expressly requires disclosure when judges or court officials or employees use AI in judicial processes, including preparing court-issued adjudicatory documents. It specifies the tool and version, purpose and extent of use, human oversight, preservation of outputs, compliance with the framework, and acceptance of responsibility. The disclosure is to follow guidelines issued by the Court.
• Confidential information. Part II.A.6 prohibits processing confidential, privileged, sensitive, or secret information through AI without express authority and strict supervision by the Court or an authorized office. Users must understand what data enters a tool, how it is processed, and who can access it.
• Risk assessment. Part II.A.7 requires a comprehensive risk assessment before implementation of an AI tool in the Judiciary. Safeguards must address inaccurate outputs, bias, security threats, and the risks of relying too heavily on automated results.

Practice implications

Parao Law practice guidance

For lawyers

If AI supplies a case citation for a draft motion, open the actual decision and check the case name, reference, quotation, and proposition before using it. If it summarizes a record, compare the summary with the original documents and retain the material used for verification.

Keep a record of the tool, version, task, outputs, and human checks. This supports accountability and any applicable disclosure. The framework expressly details disclosure by judges and court personnel; lawyers should check the Court’s implementing directions and applicable professional duties before choosing a disclosure form for a filing.

A medical chart attached to a contemplated pleading may contain sensitive health information. Uploading it to an AI tool requires an assessment of legal authority, confidentiality, and data handling. A vendor’s privacy promise alone does not supply the authority required by the framework.

For healthcare organizations and businesses

A hospital supplying records for litigation should agree with counsel on how those records will be handled, including any proposed AI processing. Identify who can access the data, whether the service retains it, and who checks the output against the original record.

The framework can inform internal governance, but its requirements concern judicial work. Clinical AI, routine business operations, and patient care require a separate assessment of the laws and professional standards governing those activities.

Scope and limits

This update explains an institutional governance framework. It does not establish a medical standard of care or authorize clinical AI. The full resolution and subsequent Court guidelines govern its application to a particular judicial task.