02/12/2025
When Justice Meets the Machine: Navigating AI’s Promise and Peril in Law
By: Atty. Arbie C. Buenaflor, JD
I recently completed an online Mandatory Continuing Legal Education (MCLE) course examining Artificial Intelligence (AI) and its increasing relevance in modern legal practice, particularly in arbitration. The course offered meaningful insights into how AI, blockchain, and decentralized systems are gradually reshaping dispute resolution. From automating research and streamlining evidence review to improving case management and providing data-driven analytics, technology is increasingly embedded in the legal process.
One highlight of the course, and what resonated with me the most, was the discussion on AI’s emerging role in arbitration, especially a 2021 case from Mexico where a court enforced an arbitral award that incorporated a ruling generated through Kleros, a blockchain-based decentralized justice platform (Ejecución de sentencia, Asunto 205/2021, Juzgado Cuarto de lo Civil, Jalisco, 28 May 2021; documented in Carrera Report, 2021).
The facts and procedure of this case were simple but unprecedented. The dispute arose from a rental estate leasing agreement between two natural persons. The agreement provided for a novel arbitration mechanism: a hybrid process. Specifically, the clause stipulated that after receiving the parties’ claims, the arbitrator would draft a Procedural Order addressed to the decentralized justice platform Kleros, which would issue a decision based on its blockchain arbitration protocol. Kleros is a decentralized application (“Dapp”) deployed on Ethereum that provides users with decentralized arbitration services. Importantly, the clause required that the arbitrator incorporate the decision received from Kleros into his arbitral award, issue it in writing, and indicate the date, place, name, and signature of the arbitrator.
On 3 November 2020, after the landlord initiated arbitration proceedings, the arbitrator followed the procedure agreed upon by the parties and issued an arbitral order titled “for referral to Kleros,” remitting the merits of the dispute to the Dapp. After running its blockchain arbitration protocol, Kleros communicated the decision of the three jurors selected for the case to the arbitrator on 23 November 2020. Four days later, the sole arbitrator rendered an arbitral award in Guadalajara, incorporating the decision from Kleros.
Interestingly, the blockchain arbitration process and its decision were detailed only in the “Arbitral Procedure” section of the award, while the “Decision” (or operative part) did not mention Kleros and appeared to attribute the ruling solely to the arbitrator, without explicitly reflecting the decision-making role played by the platform. Subsequently, the landlord sought enforcement of the arbitral award before Mexican courts. On 28 May 2021, the court issued a decision granting enforcement of the award.
Though not a purely “on-chain” arbitration, this case is significant. It demonstrates how decentralized mechanisms and conventional arbitral processes can coexist when properly structured, providing a real-world example of how technology can support—but not replace—human decision-making in legal proceedings. To fully appreciate this development, it helps to understand what Kleros is.
For us who are not techie enough, Kleros is a decentralized online dispute resolution protocol built on the Ethereum blockchain. It functions as a “crowdsourced court” for small claims and digital transactions. The platform relies on randomly selected pseudonymous jurors who stake its native token to participate. Incentives align jurors toward decisions that converge with the rational majority, and those who vote inconsistently risk losing their stake. Smart contracts automatically enforce the resulting decisions. This structure aims to provide swift, low-cost, and impartial resolution without traditional judicial intervention.
Yet Kleros also illustrates a broader truth: technology, no matter how sophisticated, remains a tool. It is engineered to address human disputes but cannot replace the complex motivations, biases, and lived realities that bring litigants to court. This realization led me to revisit my own reflections on law, technology, and the human element.
I recall attending the IBP House of Delegates in Baguio City in January 2020, during my tenure as Director of the IBP Quezon City Chapter. Then Chief Justice Diosdado Peralta discussed the Supreme Court’s revisions to the Rules of Court and observed that delays in case disposition were primarily fueled by legal practitioners, the lawyers.
While not entirely inaccurate, this observation missed a more fundamental reality: cases originate from litigants—not lawyers. A dispute begins with a human grievance, injury, or interest. Lawyers come in only after the conflict has already taken shape. To attribute systemic delay primarily to counsel is to overlook that litigation is, first and foremost, a human product.
The law, unlike mathematics or the natural sciences, is not an exact science. It cannot rely solely on algorithms, formulas, or mechanical processes because it must embody human values, principles, and judgment. Justice is not achieved merely by following rules; it requires interpreting those rules in light of context, equity, and public welfare. Law must remain human because it carries with it the responsibility to promote order, fairness, and the public good. Human wisdom, moral reasoning, and empathy are indispensable in ensuring that legal outcomes are not only legally correct but socially just. Without these human elements, even a perfectly logical ruling can fail to serve justice or society at large.
This brings me to a growing concern about the rapid adoption of AI in the justice and legal sector. Technology is designed to streamline processes, but tools inevitably shape the hand that wields them. As courts worldwide experiment with AI for research, drafting assistance, and predictive analytics, an unsettling possibility emerges: what happens if judicial workflows rely so heavily on AI that judges are reduced to inputting facts, uploading evidence, or selecting options on a screen, after which the machine produces the “recommended” outcome?
In such a model, the judge risks becoming a mere operator, and the courtroom risks transforming from a deliberative forum into a processing center. Judicial independence could be compromised if AI-generated outcomes become the default, and institutional pressures might encourage deference to machine recommendations for the sake of expediency. Accountability could become diffuse—should the coder, the judiciary, or the judge who simply clicked “approve” bear responsibility for an unjust ruling? Can a system built on human dignity survive if human judgment is marginalized?
These concerns are not speculative; they are structural risks. AI excels at speed, pattern recognition, and computation. But justice demands wisdom, empathy, moral reasoning, and context—qualities no machine possesses. Efficiency cannot replace discernment. If we allow technology to overshadow the human role in adjudication, we risk hollowing out the very essence of justice. Courts may become faster, but not necessarily fairer. As the old wisdom reminds us, “slow is fast.” Rushed justice—however technologically efficient—can amplify injustice and ultimately slow the system even more.
The MCLE course reaffirmed that AI carries extraordinary potential but equally serious dangers. It can empower lawyers, accelerate proceedings, and reduce administrative burdens, but only if it remains what it must always be: a tool for human beings, not a replacement for human judgment. As legal systems innovate—whether through decentralized models like Kleros or AI-assisted adjudication—we must stay anchored in one foundational truth: the law is a human institution, created by humans, applied by humans, and intended to serve human beings, embodying the human values necessary for order, justice, and the public good.
Fellow legal practitioners, anong tingin nyo?