AI and justice

Justice in the Philippines, as many of us know, does not move swiftly. It drips — slowly, unevenly — through clogged dockets and worn-out calendars, past reset hearings and missing case folders, often weighed down by paper, manpower shortages, and more than a few ghosts of bureaucracy. For years, we’ve accepted it with a kind of tired resignation: “Ganyan talaga.”

But now, in hushed conference rooms inside the Supreme Court’s Office of the Court Administrator, and in barangay halls experimenting with mediation support tools, something is changing. And that change is not wearing a barong or a robe — but running on algorithms and servers.

Welcome to the early stirrings of digital justice in the Philippines.

It began quietly enough. A pilot AI summarization engine was tested in two regional trial courts last year. The tool used natural language processing (NLP) to read case transcripts — hundreds of pages of affidavits, pleadings, and annotations — and distill them into digestible summaries. Not just for clerks, but for judges who often face piles of unread material days before a decision is due.

What took junior staff three days to summarize, the system could condense in under a minute. And surprisingly, with accuracy that surpassed expectations.

Soon after, the Department of Justice partnered with a local legal tech firm to run NLP-based archiving for unresolved land cases. The backlog dropped 20 percent in six months. And just last quarter, barangay officials in Quezon Province began testing a mobile app that uses AI to draft mediation agreements based on spoken accounts — transcribed in real time, translated if needed, and formatted into official forms for resolution.

The implications are staggering. For a justice system long paralyzed by backlog — over 700,000 pending cases nationwide, many of them stuck in procedural limbo — technology offers more than hope. It offers a practical lever.

But as always, with convenience comes the undercurrent of caution. Because justice is not just a matter of speed. It is also a matter of care.

When an AI summarizes a case, does it grasp the nuance of intent? When a mediation draft is generated by a voice-to-text system, can it tell the difference between sincerity and pressure? Will a judge, pressed for time, rely too heavily on an AI-generated summary instead of reading the full record?

And what happens when legal AI begins learning from biased data, reproducing errors or embedding inequities into patterns of “efficiency”?

The law, after all, is built on precedent — but justice is built on discretion.

To the credit of the judiciary, many are asking the same questions. Supreme Court Chief Justice Alexander Gesmundo has championed the Judiciary eLearning Platform and emphasized the balance between modernization and constitutional integrity. In fact, the current Digital Justice Blueprint includes safeguards against full automation, insisting on human oversight at every stage.

But the real test will not be technical. It will be cultural.

Can overworked court clerks, barangay tanods, and public attorneys learn to trust — but verify — AI tools? Will they be trained, funded, and supported, or simply handed a login screen and left to improvise? Can AI in justice remain assistive — never authoritative?

Because for all its flaws, the human justice system still asks something no machine has ever mastered: to weigh not just facts, but fairness. Not just rules, but reason.

If we are to embrace digital justice, let us do so deliberately — not for speed alone, but for clarity, accessibility, and compassion.

Because a system that decides faster is not always a system that decides better.

And when liberty, land, or livelihood hangs in the balance, we must ensure that the future of justice doesn’t just speak in code — but still listens with a conscience.