Supreme Court Framework Bans Generative AI in Philippine Courts for Now
Philippine courts are operating under a judiciary governance framework for “human-centered augmented intelligence” that keeps people responsible for judicial decisions and currently bars generative artificial intelligence from courtrooms, Senior Associate Justice Marvic Leonen said on September 23, 2026 at Microsoft Frontier Transformation Day. Leonen listed human primacy, explainability, bias audits, equal access, institutional control of digital systems and user training among core principles, adding that AI may assist research but never adjudication itself. For now the only approved AI application he named is Scriptix, a voice-to-text tool for stenographers; filings and court-produced documents should not be drafted with generative platforms unless an AI committee recommends a permitted pilot.
Filed under Policy and dated September 24, 2026, this AI4Philippines briefing treats the framework as Filipino justice-AI news distinct from prepaid consumer sachets or eGov assistants. Disclosure rules require stating the purpose and extent of any AI reliance and retaining outputs for verification, while the Philippine Judicial Academy prepares critical-use training for judges, lawyers and students.
Why it matters: Court output shapes rights and livelihoods. Hard human-oversight rules can prevent hallucinated pleadings—but only if disclosure is enforced and stenography tools stay narrowly scoped.
What it means in practice
Philippine litigators and court administrators should inventory any unofficial generative drafting habits; demand written committee guidance before pilots; assign an owner for disclosure templates; run time-boxed Scriptix quality checks without expanding to generative drafting; and prefer workflows that keep humans on findings of fact. Connect the framework to DICT’s eGovAI launch and IPOPHL’s human-creativity draft.
Caveats come first. Framework speeches are not circulars with sanctions; “research assist” boundaries can blur; and committee reviews may loosen limits later. AI4Philippines therefore presents the judiciary rules as directional policy context until published administrative orders appear.
What to watch next: first approved generative pilots, if any; how disclosure is audited; and whether trial courts get uniform templates. Readers can continue on the AI4Philippines homepage, or browse the Newsroom for additional briefings.
Bottom line: treat this update as orientation, not instruction. Philippine court AI policy is choosing human judgment over generative drafting and remains strict. Organizations that benefit most will disclose tool use, keep humans on adjudication, and refuse to confuse a keynote with finished court rules.