AI in courts

EDITORIAL: Supreme Court Justice Syed Mansoor Ali Shah’s ruling on April 11, advocating the adoption of Artificial Intelligence (AI) in judicial proceedings, provided it is anchored by a human-centered approach, has the potential to redefine the future of jurisprudence in Pakistan.

The ground-breaking pronouncement does more than set guidelines for integrating technology in the legal system: it charts a path that balances the need for efficiency, transparency and innovation with core universal values revolving around human dignity and compassion. If Justice Shah’s vision of justice delivery is adhered to, AI would work towards vital systemic improvements by expediting adjudication, improving case management systems and expanding legal knowledge, but at the same time, it would remain subordinate to judicial discretion, ethical imperatives and the irreplaceable human capacity for empathy.

While ordering the National Judicial (Policymaking) Committee and the Law and Justice Commission of Pakistan (LJCP) to formulate comprehensive AI usage guidelines for the judiciary, Justice Shah emphasised the need for safeguards ensuring that this cutting-edge technology serves only as a facilitative tool, and does not compromise judicial independence, constitutional principles or public trust.

In the Pakistani context, AI could play a seminal role in reducing the country’s massive case backlog. As of December 2024, there were 50,487 cases pending in the Supreme Court, while the LJCP had reported in June last year that there were over 2.2 million pending cases nationwide, including 1.8 million languishing in the district courts.

Given the massive quantum of cases awaiting adjudication across the judicial system, it goes without saying that such delays undermine the rights of litigants, compromise the very essence of justice delivery and erode public trust in the judiciary, creating a crisis that demands urgent redress.

AI could prove transformational here by automating routine tasks, optimising docket management, accelerating legal research and enabling data-driven case prioritisation, and thus significantly reducing case processing times, allocating judicial resources more effectively, and ultimately improving access to timely justice – which after all, is the core purpose of any justice system worth its name.

At the same time, as very aptly put by Justice Shah, “the rule of law must always prevail over the rule of data”.

The ultimate authority to interpret evolving legal and moral standards cannot be handed over to technology as it does not possess human judgment, ethical reasoning, the lived experience of judges or the capacity for mercy essential to justice. Hence the ruling very rightly mandates that AI algorithms serve rather than supplant judges, ensuring technology never eclipses an individual judge’s moral compass, compassion or societal values.

This is backed by both global experience and research. There have been numerous instances worldwide of the use of AI tools in criminal justice leading to wrongful arrests due to biased predictive policing algorithms that perpetuate historical biases and prejudices against marginalised communities. For instance, Amnesty International has found UK police using predictive algorithms that racially profile communities, with 75% of police forces in that country employing this discriminatory system.

A 2024 Unesco survey has also shown how AI’s use in judicial processes carries significant risks of human rights abuses pertaining to privacy, freedom of expression and non-discrimination. Also of significance is the US Supreme Court’s warning against AI-generated “hallucinations”, where AI chatbots produce non-existent case citations that have known to lead lawyers to file briefs with fabricated references, jeopardising the integrity of legal processes.

So, alongside adopting Justice Shah’s human-centric approach, there is also a clear need to craft and implement effective AI governance protocols, and establish training programmes geared towards ensuring responsible AI use in judicial activities. Ultimately, AI should assist the justice system – not replace the judges who give it meaning.

Copyright Business Recorder, 2025