Equity’s Exile: Party Autonomy in Blockchain Arbitration – Part I
- Introduction
Blockchain arbitration is emerging as a new frontier in dispute resolution. Smart contracts can now trigger instant outcomes through decentralized jurors or AI systems, with decisions executed automatically. The appeal is obvious because of faster resolution, lower costs, and fewer intermediaries. Yet this technological leap collides with arbitration’s evolving recognition of ‘equity.’
Over decades, arbitration has increasingly embraced human judgment, consent-based discretion, and fairness beyond fixed rules. Blockchain systems narrow that space. This creates a paradox. Parties expect efficient and fair dispute resolution yet choose structures that make such judgment impossible by design. This essay examines that dilemma and explores how equity might still be preserved within code.
- Equity in Arbitration: The Human Foundation
Equity in arbitration refers to consent-based power to decide beyond strict law. Known as ex aequo et bono or acting as an amiable compositeur, equity allows arbitrators to consider context, hardship, intent, and proportionality, where strict rules might undermine justice. This is not indulgence, but a recognition that fairness cannot always be reduced to mechanical obedience.
Equity in arbitration must be explicitly authorized. Section 28(2) of India’s Arbitration and Conciliation Act allows tribunals to decide ex aequo et bono only with explicit party consent. This is mirrored in Article 28(3) of the UNCITRAL Model Law and Section 46 of the English Arbitration Act 1996. Major institutional rules like ICC Article 21, LCIA Article 14, and ICDR Article 16(3) have similar requirements. This reflects a shift toward favouring fairness over strict legalism.
Following the 2015 amendment to section 28, Indian tribunals are required to take contract terms “into account,” rather than follow them mechanically. This marks a growing trend toward interpretative flexibility. In Batliboi Environmental Engineers v. Hindustan Petroleum (2023), the Supreme Court acknowledged that arbitrators may invoke fairness and equity in appropriate cases, provided awards remain reasoned and procedurally disciplined. Courts reviewing awards under the public policy exception of the New York Convention (Article V(2)(b)) have similarly accepted equitable outcomes, provided they remain aligned with the contract’s spirit.
Importantly, equity is inseparable from deliberation, reasoning, and dissent. Justice requires not only just outcomes but justification. Reasoned awards, minority opinions, and principled divergence are not a flaw but a feature of adjudication. Arbitration’s structure, which emphasizes party choice, reasoned decisions, and limited judicial review, rests on the belief that justice is determined actively, rather than calculated passively. Dissent plays a crucial role. Courts have recognized dissenting opinions as safeguards against arbitral error, and they consider them during judicial review under Section 34 of the Arbitration Act. Well-reasoned dissent enhances tribunal quality by urging majority arbitrators to tackle issues thoroughly and thoughtfully.
This fundamentally human groundwork now faces a challenge from the automated nature of blockchain arbitration.
III. Blockchain Arbitration: Human Judgment to Algorithmic Determinism
Blockchain is an immutable distributed ledger without a central authority. Once recorded, transactions cannot be reversed. This creates structural challenges when disputes arise from smart contracts. While tribunals can issue reasoned decisions, they cannot command code to undo execution. Law can declare, but code simply executes. This friction is pushing dispute resolution onto the blockchain itself. As the change unfolds, the system shifts from human judgment to algorithmic convergence.
Off-chain arbitration still follows the familiar path. Institutions such as ICC and SIAC rely on human interpretation of contracts and of law and equity, as appropriate. The New York Convention makes the awards contestable and enforceable. This weakness appears at enforcement. Courts cannot undo on-chain transfers or easily identify pseudonymous actors. Equity survives in doctrine but struggles in execution.
On-chain arbitration rewrites the architecture. Here, dispute resolution is integrated with the blockchain, and the results are automatically implemented. Models differ but share the same logic. Some use decentralized juror pools that stake tokens and vote. Others depend on Oracle networks that feed verified external data for the decision. More advanced systems experiment with AI-driven decision engines trained on past disputes. Decisions become faster, tighter, and self-executing.
What disappears is real-time human judgment. There is no deliberation, no moral hesitation, no dissent. The system is designed to converge, not to reflect. Code values predictability and speed, whereas equity requires discretion and context. That quality has no place in automated execution.
This results in the central paradox. Equity is not a formula but a corrective instinct that allows decision-makers to deviate when strict application of the law would cause injustice. Algorithms are not capable of this. Algorithms optimize both the input and output, but not for justice or the prevention of unfairness. In pure on-chain systems, the marginalization of equity is not a random phenomenon, but is a designed-out structure.
This tension plays out most vividly in specific systems.
A. The Kleros Paradox: Convergence over Conscience
Kleros makes the tension between discretion and convergence a live experiment. It operationalizes Schelling’s game theory, which holds that people acting individually are more likely to converge on salient “focal points” when they cannot communicate. Jurors are randomly selected, stake PNK tokens, and vote for the outcome they believe the majority will choose, not what they personally believe is right. The ones who agree with the final majority retain their interest and receive a portion of the tokens from those who disagree, while those who disagree lose their interest.
For clear-cut disputes, this works well. Reported coherence rates are high, and automated execution delivers speed and certainty. Equity matters most in uncertain cases, hardship, and grey areas with multiple plausible outcomes. That’s where the challenge arises. Once there is no clear object of interest, the jurors will face a difficult decision: either vote their conscience and lose money or vote based on what they believe others will vote. The system eventually becomes selective on conformity, not courage.
Traditional arbitration treats equity as a controlled departure from strict legality. This has been gradually acknowledged by Indian courts, which give arbitrators a circumscribed freedom to give precedence to substantive fairness rather than mechanical legality. Kleros inverts that logic. A juror who votes equitably against the literal outcome of a smart contract does so at personal economic risk. Dissent is not recorded as a principled minority view but is punished as incoherence. It does not result in deliberation to arrive at justice but convergence to pass as legitimacy.
B. AI Arbitrators: Equity as Algorithm
The sphere of arbitration has already been penetrated by artificial intelligence. Analytics tools increasingly predict outcomes based on past cases, while document review and evidence handling are automated. These systems enhance efficiency, but the final decision is made by humans.
The more radical suggestion is for fully autonomous AI arbitrators, trained on large corpora of contracts, statutes, and past decisions, to issue binding awards. The CIArb 2025 Guidelines caution that these may jeopardize the enforceability of the award due to their imprecise outputs, demanding human intervention. In such a system, equity is no longer a matter of conscience but a pattern in historical data. Justice becomes what the model has most often seen before.
This raises issues of on-chain systems and exacerbates them. Equity is what keeps rigid legality from becoming blind rule-application in the interest of fairness. A model can give a rough idea of how the courts have performed; however, it cannot tell why a deviation is now warranted. Its arguments are mostly not transparent, and its results, though true, are hard to dispute. The same is supported by Article 18 of the UNCITRAL Model Law, as well as the enforcement framework in the New York Convention, which presupposes awards that parties may comprehend, review, and appeal. A system that optimizes predictive power and speed but cannot account for itself in human terms risks reproducing the same paradox: efficiency scales upward, but equity thins out.
Conclusion
Part I established the doctrinal foundations of this conflict, framing equity not merely as a discretionary tool but as a consent-based corrective. By examining UNCITRAL Model Law, and its parallel in Section 28(2) of India’s Arbitration Act, it demonstrated that equity is inextricably linked to human deliberation and principled dissent. Conversely, an analysis of token-incentivized models like Kleros reveals a structural design intended to eliminate these very features. This shift represents more than a gain in efficiency, it is largely a displacement of equitable judgment. Part II examines the “consent paradox” inherent in these decentralized frameworks and outlines a path forward via hybrid design.
*Shivanshi and Arzoo are Second year students at Hidayatullah National Law University, Raipur.