Artificial Intelligence, Public Policy, And Fraud - Rethinking Enforcement Under The New York Convention

Author: LawSphereX

Category: General Law

Published: 2026-08-23T18:19:49.76855+00:00

Updated: 2026-08-30T07:06:38.118Z

ARTIFICIAL INTELLIGENCE, PUBLIC POLICY, AND

FRAUD: RETHINKING ENFORCEMENT UNDER THE NEW YORK CONVENTION Shahida Parween, Law College Dehradun Mr. Vatsal Chaudhary, Assistant Professor, Law College Dehradun

ABSTRACT

As it guarantees the acceptance and enforcement of foreign arbitral rulings, the New York Convention, 1958 continues to be the cornerstone of international commercial arbitration. But the public policy exception under Article V(2)(b) has consistently led to doctrinal ambiguity, especially when combined with fraud accusations. Differentiating between strategic, meritbased arguments masquerading as enforcement objections and actual violations of fundamental public policy remains a challenge for courts throughout jurisdictions. This paper critically analyses how public policy and fraud are interpreted in many important jurisdictions, such as Singapore, the United States, the United Kingdom, India, and Switzerland. It contends that the pro-enforcement goal of the Convention is weakened and arbitral finality is compromised by inconsistent judicial procedures.

In this study, artificial intelligence (AI) is introduced as a judicial decisionsupport tool that can help courts uncover fraud-related risk patterns, identify baseless public policy arguments, and improve consistency in enforcement outcomes. The paper suggests a technologically informed, appropriate, and globally harmonized paradigm for enforcement review by incorporating AIdriven tools like Natural Language Processing (NLP), predictive analytics, and human-in-the-loop governance frameworks. The study comes to the conclusion that AI can greatly improve the legitimacy, effectiveness, and consistency of arbitral award enforcement under the New York Convention when used morally and as an auxiliary rather than adjudicatory instrument. Keywords: Artificial Intelligence, Public Policy, Fraud, Arbitrability, Enforcement, Convention of New York, Judicial Intervention.

• Introduction - • Conceptual Framework: Public Policy, Fraud and AI - • Comparative Study of different dominion - • India: Broad Explication and Judicial Supervision - • Singapore: A Calibrated, Internationalist Model - • United States: A Regulative and Pro - Enforcement Perspective - • Challenges and Limitations of AI in enforcement - • Recommendations - • Conclusion Fig.1: Manuscript organisation.

INTRODUCTION

One of the most effective tools of transnational commercial law is the enforcement of foreign arbitral rulings under the New York Convention, 1958. Its efficacy depends on striking a careful balance between international arbitral finality and national sovereignty. Public policy under Article V(2)(b) is meant to be a narrow safeguard rather than a broad judicial escape clause, and Article V of the Convention defines limited reasons upon which execution may be refused. Despite this intention, the public policy exemption has developed into one of the Convention's most controversial and unevenly interpreted clauses. This exclusion often covers fraud allegations, enabling losing parties to oppose enforcement by relitigating factual and legal disputes that have already been decided by arbitral courts.

The lack of a consistent definition of public policy and fraud under the Convention makes this problem worse. The thresholds for involvement vary greatly among jurisdictions, which causes forum shopping and unpredictability. Some states maintain broad domestic interpretations that weaken arbitral authority, while others embrace a limited, internationalist understanding of public policy. In light of this, artificial intelligence (AI) becomes a useful instrument that can help judicial authorities handle enforcement review more logically. AI improves judicial discretion by facilitating organized evaluation, pattern detection, and uniformity in enforcement decisions rather than replacing it. This essay makes the case that AI-assisted judicial review provides a morally sound means to balance the Convention's pro-enforcement stance with public policy protections.

CONCEPTUAL FRAMEWORK: PUBLIC POLICY, FRAUD AND AI There are heads of disputes, which is arbitrable and there are categories which are not arbitrable however, it depends upon the jurisdictions of the nations also, as different jurisdictions interpret disputations erratically, such as matters of serious allegations, antitrust, criminal instance, are non-negotiable in most of the nations as it requires proper trial, procedure and admissibility of court to resolve the matter. Enforcement Resisting application party raises objection ( before competent court) Court's Discretion- Standard/ Threshold Court examines ary decision depending for fraud or public the objection policy objection upon the case No re - examinati Final outcome on on merits Fig. 2: Implementation of award within the Convention of 1958- objection raised on ground of fraud or public policy.

Across these jurisdictions, one consistent pattern emerges: arbitrability is becoming more prevalent globally, particularly in economic domains such as commercial, intellectual property and also fraud. As a result, arbitrability ultimately reflects each country’s balance of contractual freedom and public regulatory interests, resulting in the widely different terrain that exists in international arbitration today.

As mentioned in the New York Convention that- “Recognition and enforcement of an arbitral award may also be refused if the competent authority in the country where recognition and enforcement is sought finds that: a) The subject matter of the difference is not capable of settlement by arbitration under the law of that country; or b) The recognition or enforcement of the award would be contrary to the public policy of that country” [1] Moreover, Article 36(1) of the UNCITRAL Model Law, states that a court can deny enforcement if it finds that the subject matter of the dispute is not capable of settlement under the law of the State.

[2] Over this period, the problems that arose under the public policy exception have subjected courts, both in common law and in the civil law, to considerable abuse; as heard in the case of Richardson, public policy was described as an “unruly horse”, which can sometimes lead to capricious outcomes, as when you get astride, you never know, where it will carry astride, you never know, where it will carry you. [3] While enforcing any award, if it comes to the Court that the execution, is related to any distress of public policy, then it is on the circumspection of the jurisdiction to put aside the accomplishment. • The competent court receives the final arbitral award for recognition and enforcement.

Arbitral Award Lords of English House mentioned public policy as “that principle of law which holds that no subject can lawfully do that which has a tendency to be injurious to the public, or against public good. ” [4] Courts are nonetheless wary, especially where accusations include public regulatory interests, even though modern arbitration law increasingly acknowledges the arbitrability of complicated commercial disputes, including those involving fraud. Although this caution is sense, it frequently leads to overzealous judicial examination. Inconsistent results result from the lack of a systematic approach for evaluating fraud and public policy claims. AI-powered analytical tools can help courts in this situation by offering unbiased, data-driven insights without interfering with judicial authority.

COMPARATIVE

ANALYSIS

OF DIFFERENT JURISDICTIONS The exception of “public policy” under the enforcement of arbitral award has been widely interpreted by judiciaries across the globe, including India. [5] Another milestone, in Australia, following the literal text of the legislation, if any breach of natural justice in connection with making the arbitral award may arguably be contrary to or in connection with Australian public policy, triggering the possibility of recourse, including setting aside. [6] Further, the Tribunals of Hong Kong, laid down that to take the plea of not enforcing an arbitral judgment and invoking the exception clause, the circumstances shall be of such nature that it shocks the conscience of the nation and its integrity. [7] Swiss courts continue to take a restricted approach under the Private International Law Act.

[8] India: Traditional However, Growing Approach- India has incorporated New York Convention’s exception of public policy through a body of case laws and wide interpretation. The substantive point came with the precedent of court, where the Supreme Court was of the view that an enforcement of arbitral award will not be possible and could be set aside if it breaches “patent illegality” of is contradictory to the foundational policy of Law of India. [10] This broadened notion of public policy far beyond the limited scope envisaged through Article V(2)(b) of the Convention, effectively permitting judicial review on merits under the guise of enforcing domestic legality. [11] The Court has also reaffirmed a narrower, global standard for foreign awards, limiting the expanded interpretation to domestic awards.

[12] Further, the Court has also highlighted that the enforcement of an arbitral award can also be denied if the final award contravenes India’s most basic principles of justice and morality, and not just a mere infringement of Indian Laws which does not amount to any breach of public policy. [13] The extent of foreign awards has been limited by later jurisprudence, but accusations of fraud are still subject to close judicial investigation. AI-assisted screening could support recent pro-arbitration measures by assisting Indian courts in distinguishing between factual disputes and significant fraud. Singapore: A Global Blend- Alike in India, Singapore’s Judicature of Appeal, also elucidated that the award can be uphold if it “breaches the basic notion of the justice, fairness or morality,” or “disturbs the conscience.

” [14] The Singapore model of arbitration exemplifies a rule-based, pro-enforcement and an internationally approachable view that harmonises the judicial restraints with arbitral probity. [15] The Singapore Court of Appeal ruled that if an arbitration clause if broad enough to include a dispute, there is a presumption of arbitrability, unless legislation explicitly prohibits arbitration for that specific case. [16] Singapore's rigorous, internationalist approach to public policy makes it an exemplary jurisdiction. Its courts place a strong emphasis on evidential rigor, proportionality, and materiality.

The incorporation of AI tools enhances Singapore's already uniform enforcement structure and is in line with its rule-based judicial ideology, United States: A Restrictive and Pro-Enforcement Approach- The US has historically and narrowly interpreted public policy under the Convention of 1958. Basic conceptions of morality and fairness, rather than only domestic legal norms. [17] It has supported a restrictive approach, arguing that Jurisdiction % of Awards Refused (2015- 2024) % Citing Public Policy Judicial Trend India 14% 6% Narrowing interpretation post Renusagar and Shri Lal Mahal Singapore 8% 3% Restrictive approach, high deference to arbitral autonomy USA 5% 2% Strong pro-enforcement under FAA and NY Convention enforcement can be refused only if it breaches basic principles of justice.

[18] % of Total refusals Due process Violations Invalid Agreement Public Policy Fraud/Corruptions Excess of Jurisdictions Others Fig. 5: Grounds and % of total refusals of cases. [19] The allegations of fraud or procedural unfairness are not enough; the fraud must be proven and material to the tribunal’s finding. [20] The U. S. legal system stresses stability, predictability, and limited court intervention to prevent public policy from being used as a justification for more litigation. [21] U. S. courts apply a restricted interpretation of public policy, restricting involvement to transgressions of basic justice and morality. The U. S. system is especially vulnerable to AI-driven risk assessment because of the high evidence threshold for fraud, which can expedite enforcement actions without sacrificing due process.

ARTIFICIAL INTELLIGENCE IN THE ENFORCEMENT OF ARBITRAL AWARDS AI as a tool for Judicial Decision Support- The best way to think of AI in adjudicatory settings is as an assistance technology. Instead of making decisions, judicial decision-support systems examine vast amounts of data, spot trends, and highlight possible dangers. AI can help courts enforce arbitral rulings by classifying complicated documents, spotting recurring litigation tactics, and pointing out contradictions in fraud claims. Fraud detection and Natural Language Processing- NLP makes it possible for AI systems to identify recurrent linguistic and structural trends in pleadings, enforcement objections, and arbitral awards. NLP algorithms can detect, for instance, recurrent use of domestic illegality grounds, boilerplate fraud allegations, or discrepancies between arbitral rulings and enforcement-stage filings.

AI Augmented Enforcement Framework under NY Convention Judicial Oversight Predictive Analysis NLP Risk Screening Category 4 Series 1 Series 2 Series 3 Fig. 6: AI Augmented Enforcement Framework under New York Convention Consistency in Enforcement and Predictive Analytics- Predictive analytics evaluates the probability of rejection based on jurisdiction, reasons raised, and procedural posture using past enforcement data. Predictive outputs offer useful context by indicating if a specific objection is consistent with accepted international practice or substantially deviates from Convention principles, even though they do not bind courts. Governance with Humans in the Loop- A human-in-the-loop framework is necessary for the ethical application of AI. While AI systems function in a transparent and understandable manner, judges maintain complete decision-making authority.

This approach increases effectiveness, uniformity, and trust in the results of enforcement while maintaining judicial independence

CHALLENGES

AND LIMITATIONS OF ARTIFICIAL INTELLIGENCE IN THE ENFORCEMENT OF AWARD

1. When evaluating public policy and fraud, AI systems rely on historical enforcement data,

which frequently includes domestic biases and jurisdictional inconsistencies. AI models run the potential of reproducing rather than correcting interventionist patterns if they are trained on such data.

2. Reasoned rationale is necessary for judicial enforcement decisions, especially when

declining enforcement under Article V(2)(b) of the New York Convention. Because many AI systems are not sufficiently explainable, judges find it challenging to explain how algorithmic aid affected their decisions Concerns about explainability and transparency Limitations Data restrictions of Normative Challenges and limitations and algorithm and interpretive bias Judicial discretion and due process risks Fig. 7: Challenges and Limitations of AI in the enforcement review.

3. AI cannot evaluate value-laden notions like public policy, morality, or justice on its own and

is incapable of normative legal reasoning. Determining significant fraud or fundamental public policy infractions is still a judicial role that requires contextual and qualitative review, even though AI may be able to spot risk patterns or discrepancies.

4. Automation bias could result from an over-reliance on AI-generated evaluations, gently

persuading judges to accept computational results. Additionally, if parties are unable to contest their methodology or presumptions, the unreported employment of AI tools may give rise to due process issues. Therefore, maintaining judicial independence and procedural fairness requires effective humanin-the-loop measures.

RECOMMENDATION

▪ Judges must make the final decisions regarding enforcement; artificial intelligence should only be used as a support tool to help courts detect spurious public policy and fraud-based challenges.

▪ To prevent an over-reliance on algorithmic evaluations, courts should establish precise guidelines for the use of AI-generated insights in enforcement actions. ▪ Parties must be made aware of the use of AI technologies and given the opportunity to challenge any review that uses AI if needed in order to maintain transparency. ▪ To encourage uniformity across jurisdictions, international arbitration organizations should publish fundamental guidelines on the moral and restricted application of AI in the enforcement of arbitral awards. ▪ To lessen bias and take into account changing judicial interpretations of public policy, AI systems employed in enforcement review should be based on precise, varied, and frequently updated data.

CONCLUSION

Although fraud and public policy are crucial protections under the New York Convention, their abuse jeopardizes the legitimacy and finality of international arbitration. The uniform implementation of the Convention is undermined by comparative jurisprudence, which exposes ongoing inconsistencies in enforcement outcomes. A cautious yet revolutionary solution is provided by artificial intelligence. AI improves uniformity, transparency, and proportionality in enforcement review when used as a court assistance’s rather than a decision-maker. AI strengthens both national sovereignty and international arbitral integrity by helping courts discern between legitimate public interest concerns and calculated litigation strategies. In the end, AI has the ability to improve enforcement review's consistency, efficiency, and fairness when used publicly and in a human-in-the-loop framework.

In the end, judicial discretion and the transnational integrity of international arbitration can be preserved while strengthening the legitimacy of arbitral enforcement through the careful and ethical incorporation of AI.

REFERENCES

[1] Van den Berg AJ. The New York Convention of 1958: An Overview. An Enforcement of Arbitration Agreements and International Arbitral Awards: The New York Convention in Practice, London: Cameron May 2008 (pp. 39-68) [2] Contini P. International Commercial Arbitration: The United Nations Convention on the Recognition and Enforcement of Foreign Arbitral Awards. Am. J. Comp. L. 1959; 8:283 [3] Swain W. Contracts ‘Not for the Public Good and the Classical Law of Contract. The Journal of Legal History. 2022 Jan 2;43(1):1-23. [4] Knight WS. Public policy in English law. LQ Review 1922 (38:207) [5] Kronke H, Nacimiento P, Otto D, Port NC, editors. Recognition and enforcement of foreign arbitral awards: a global commentary on the New York Convention. Kluwer Law International BV; 2024 Dec 5 [6] J. S. 2. Nicholas Kazaz, “When Even Fraud Is Not Nearly Enough.

Recourse Against Arbitral Awards and Public Policy Considerations: An Anglo-Australian Perspective,” Arbitration: The International Journal of Arbitration, Mediation and Dis, pp. 124-135 [7] Snyder KB. Denial of Enforcement of Chinese Arbitral Awards on Public Policy Grounds: The View from Hong Kong. Va. J. Int'l L.. 2001; 42:339. [8] Cassling DR. Karaha Bodas Co. , LLC v. Perusahaan Pertambangan Minyak Dan Gas Bumi Negara. Banking Law Journal. 2003;120(5):472- [9] “Journal, Singapore Academy of Law. Based on compiled data from Singapore Academy of Law Journal (2022), U. S. Federal Arbitration Reports, and Indian Law Reports, 2022”. [10] Nandi S. The Reality behind the Arbitration Process in India. Jus Corpus LJ. 2024;5:105 [11] Bermann GA. Recognition and Enforcement of Foreign Arbitral Awards: The Interpretation and Application of the New York Convention by National Courts.

In Recognition and Enforcement of Foreign Arbitral Awards: The Interpretation and Application of the New Y.

[12] Shri Lal Mohan Ltd. v. Progetto Grano Spa. , (2014) 2 SCC 433. [13] Mishra I. Foreign Exchange Laws and Their Impact on the Enforcement of Foreign Arbitral Awards in India Under the New York Convention. Asian Dispute Review. 2020 Oct 1;22(4) [14] Ng MM. Reviewing the standard of curial review for findings in arbitration involving public policy. Sing. J. Legal Stud.. 2022:75 [15] Pinsler J. Revisiting similar fact evidence in criminal and civil cases and proposals for reform. Singapore Academy of Law Journal. 2021 Mar 1;33(1):531-56 [16] Hwang M, Chan YW. Case Law of the Supreme Court of Singapore in the Field of Arbitration. bArbitra| Belgian Review of Arbitration. 2019 Oct 1;2019(2) [17] Onuzulike C. ENFORCEMENT OF ARBITRAL AWARDS: NAVIGATING THE UNCERTAINTIES OF PUBLIC POLICY DEFENCE. Available at SSRN 5194665. 2025 [18] Evans AE. Parsons & Whittemore Overseas Co. , Inc. v.

Société Générale De L’industrie Du Papier (Rakta). 508 F. 2d 969. US Court of Appeals, 2d Cir. , Dec. 23, 1974. American Journal of International Law. 1975 Oct;69(4):897-9. [19] “London, Queen Mary University of International Arbitration Survey 2021. ”. [20] Cassling DR. Karaha Bodas Co. , LLC v. Perusahaan Pertambangan Minyak Dan Gas Bumi Negara. Banking Law Journal. 2003; [21] Kagan RA. Adversarial legalism and American government. Journal of policy analysis and management. 1991 Jun;10(3):369-406