Philippine courts have adopted a framework for artificial intelligence (AI) that keeps people responsible for judicial decisions and requires users to disclose when they use AI in court processes.
Speaking at Microsoft Frontier Transformation Day, Supreme Court Senior Associate Justice Marvic Leonen said the judiciary adopted its governance framework on “human-centered augmented intelligence” this year. It sets principles for using AI while the courts examine which tools, if any, should be allowed for particular tasks.
“First, human primacy, which means that AI assists in research but never in adjudication itself,” he said.
Leonen also identified explainability, bias audits, equal access, institutional control of digital systems, and training for users among the framework’s principles. He said the courts must consider the effect of AI on people as well as its implications for cybersecurity and data privacy.
“The closer technology comes to affecting rights, opportunities, livelihoods, access to essential services, and bringing about social justice, the more important human oversight becomes,” he said.
The framework requires disclosure of AI use in judicial processes, including the purpose of each use and the extent of reliance on the tool, Leonen said. AI output must also be kept for reference or verification. He said the court determines how much AI use is allowable.
Leonen described strict limits on what can be used in the courts now. The only AI application he identified as approved for the court system was Scriptix, a voice-to-text transcription tool used by court stenographers.
“For now, no generative artificial intelligence may be used in our courts,” he said.
He added that documents filed with the courts or produced by them should not be drafted using generative AI platforms unless the court permits it on the recommendation of an AI committee. Leonen said the committee is reviewing applications and the possibility of a pilot program.
Leonen drew a distinction between AI and the judiciary’s wider shift to digital systems. Electronic filing, digital records, and online case management change how documents and information move through the courts. AI, he said, can summarize, recommend, or generate material that may affect work traditionally done by people.
“Relying on an AI-generated output in making a consequential decision presents the greatest risk and responsibility which we are not yet willing to conceive,” he said.
The judiciary is also preparing training on the critical use of AI through the Philippine Judicial Academy, Leonen said. He cited lectures for judges, lawyers, and law students, with plans to reach more parts of the country.
Leonen said the framework’s safeguards are meant to guide decisions about AI as the technology develops. He described the question facing institutions this way: “Can they use artificial intelligence without surrendering human judgment?”
