Artificial Intelligence in International Mediation and Arbitration: A Global Panorama

Abstract

This article analyzes the disruptive transformation of Alternative Dispute Resolution (ADR) through the integration of Artificial Intelligence (AI). Using qualitative and comparative analysis, it examines the transition from administrative assistance tools to systems featuring triage, decision support, and automated adjudication functions. The study evaluates the "soft law" normative response (SVAMC, CIArb, IBA) in the absence of binding state legislative frameworks, contrasting implementation models in Asia, Latin America, and emerging markets. Finally, critical ethical challenges are identified —particularly the "black box problem," algorithmic bias, and the preservation of confidentiality— proposing a "procedural legitimacy test" as a standard to ensure that technological efficiency does not undermine due process and party autonomy in contemporary private justice.

Keywords: Artificial Intelligence, International Arbitration, Mediation, ODR, Soft Law, Algorithmic Ethics, Digital Justice.

I. Introduction

The disruption of artificial intelligence in alternative dispute resolution (ADR)

Alternative Dispute Resolution (ADR) is undergoing an unprecedented transformation, driven by the convergence of economic globalization and technological acceleration. While the digitization of processes —initially under the concept of Online Dispute Resolution (ODR)— began as a practical response to conflicts arising in digital markets, the current integration of Artificial Intelligence (AI) has shifted the debate from mere communication facilitation toward a profound normative influence on the adjudication and mediation process (9).

AI is no longer perceived as a decorative or merely administrative layer within legal technology; today it acts as a functional element in case triage, evidence management, settlement design, and decision support (9). This paradigm shift means that technology has ceased to be a neutral channel and has become an actor capable of shaping the parties' perception of what is legally realistic and feasible (9). Today, generative AI tools and large language models (LLMs) are revolutionizing international arbitration and mediation, offering promises of efficiency and cost reduction that are particularly attractive in sectors with high case volumes or critical financial pressures (2, 11).

However, this "dispute revolution" is not without tensions (18). The rapid emergence of AI raises legal questions about accuracy, data control, confidentiality, and transparency, even calling into question the future viability of "robot arbitrators" (12). AI's capacity to process complex information and detect patterns in vast datasets (4) collides with the fear of losing ethical judgment and human empathy, elements intrinsic to the work of neutral third parties (7). In this sense, the international community faces the challenge of balancing the benefits of automation with preserving the integrity of the arbitral and mediation process (4).

Research questions and objectives of the article

This paper sets out to critically analyze how AI is reshaping the practice of mediation and arbitration on a global scale. To this end, the following core research questions are posed:

• To what extent is AI adoption in ADR driven by a genuine need for efficiency versus the risks of dehumanizing the justice process?

• Are the current soft law frameworks and institutional guidelines sufficient to mitigate ethical challenges such as algorithmic bias and lack of transparency?

• How do jurisdictional implementation models vary between advanced economies and emerging markets, and what comparative lessons can be drawn from these divergences?

• What is the real impact of AI tools on the validity and enforceability of awards and settlement agreements under international conventions?

The main objective of this article is to provide a comprehensive and technical overview of the state of the art of AI in ADR. It seeks to break down not only the operational capabilities of new platforms, but also the procedural integrity risks that emerge when a machine participates in framing the parties' legal options (9). The article aims to serve as a critical guide for academics and practitioners navigating this "shifting terrain" (17).

Methodology and structure of the work

The research is grounded in a comparative doctrinal and institutional analysis, employing a robust qualitative approach (9). A systematic review of academic literature published through 2025 was conducted, drawing on international databases to capture technological applications and their ethical-legal implications (7). The study critically relies on the normative guidelines of leading bodies such as the Chartered Institute of Arbitrators (CIArb) (4), the Silicon Valley Arbitration and Mediation Center (SVAMC) (5), and the International Bar Association (IBA) (6), as well as empirical adoption data and market analysis (3, 19).

The work is structured in seven sections. Following this introduction, Section II examines the global adoption landscape, analyzing legaltech market trends and the regional disparity between technology leaders and laggards. Section III delves into the soft law analysis, evaluating the institutional guidelines that now define the due diligence standard for AI use. Section IV presents a detailed jurisdictional study, contrasting China's integrated state model with Singapore's modular approach and emerging perspectives in Latin America and Africa (9, 11, 13, 14). Section V describes technology in action, from automated arbitration platforms to AI-powered mediator search tools from the American Arbitration Association (2, 11). Section VI addresses critical ethical and legal challenges: the "black box problem," bias, and confidentiality (4, 5). Finally, Section VII offers conclusions and forward-looking perspectives on the evolution of digital justice.

This article argues that the success of AI in ADR should not be measured solely by speed or cost savings, but by a "procedural legitimacy" test that ensures party autonomy and rights are not diminished in the pursuit of technological efficiency (9).

II. Global Panorama: Adoption and Trends

2.1. Empirical data and the realities of adoption in 2025

The adoption of artificial intelligence in international arbitration has ceased to be a theoretical projection and has become a quantifiable statistical reality. The current landscape reflects a transition from traditional technology-assisted document review tools toward a deep integration of generative AI into the workflow of arbitrators and lawyers (2). According to sector trend reports, 2024 marked a turning point in the expansion of these technologies, driven by growing demand for greater oversight, security, and transparency in data handling (2).

Globally, the use of generative AI tools among the working-age population reached 17.8% in the first quarter of 2026 (19). However, in the dispute resolution sector, this figure is significantly higher in leading jurisdictions. The growth of the dispute resolution solutions market is intrinsically linked to digitization: the global market for these solutions, valued at $11.1 billion in 2025, is estimated to reach $19.8 billion by 2034, with a compound annual growth rate (CAGR) of 6.7% (3). This exponential growth responds to organizations' need to resolve disputes quickly, fairly, and cost-effectively in a highly interconnected business environment (3).

2.2. Market trends and growth prospects

The LegalTech and ADR solutions market is undergoing structural transformation. The most prominent trends for the 2025–2034 period include:

• Negotiation automation: there is a growing reliance on AI-powered tools to analyze conflict patterns and predict resolutions, enabling parties to find common ground without the need for in-person meetings (3).

• Hybrid models: there is rising demand for models that combine traditional mediation with digital technologies, integrating cloud-based solutions and mobile applications to facilitate remote, flexible resolutions (3).

• LegalTech investment: the flow of capital into this sector is massive. In 2024, global venture capital investment in LegalTech reached $2.2 billion, with AI-based solutions capturing the bulk of this funding (16). In fast-growing markets such as India, the LegalTech sector is already valued at $1.3 billion (16).

• Functional specialization: AI has evolved from a simple search tool into performing complex tasks such as digesting interview memoranda, identifying contradictions in testimony, and predictive analysis of outcomes based on historical data (2).

Despite the optimism, the market faces challenges such as resistance to change in organizations accustomed to traditional practices, high initial implementation costs, and persistent concerns over data security and privacy on ODR platforms (3).

2.3. Comparative regional analysis: leaders and laggards

AI adoption in ADR is not uniform, revealing a growing technological gap between the Global North and the Global South (19).

• The leaders: UAE, Singapore, and China: the United Arab Emirates (UAE) leads the world in AI adoption, with 70.1% of its working population using generative AI tools as of early 2026 (19). This leadership is the result of a state strategy that began in 2017 and has integrated AI into legislation and public administration at an unprecedented scale (19). Singapore follows closely with a "supervised modularity" model, where e-Negotiation and eMediation court platforms are fully integrated into the judicial system under a mature and pragmatic AI governance framework (9). China, for its part, has implemented the most ambitious "smart courts" model, where AI assists across the entire judicial cycle, from blockchain-based evidence submission to judgment-drafting assistance (9).

• Latin America and the case of Chile: Chile has established itself as the regional leader in AI development, topping the Latin American Artificial Intelligence Index (ILIA) (11). With a projected sector growth rate of 33% annually, Chile outpaces larger economies such as Brazil and Mexico in terms of per capita AI spending (11). Institutions such as CAM Santiago have pioneered online resolution platforms for consumer disputes, using algorithms to propose automated solutions (11).

• Emerging markets and challenges: India, Pakistan, and Nigeria: India shows exceptional dynamism; platforms such as Presolv360 already report having resolved more than a million disputes with cost savings of up to 65% compared to traditional litigation (16). However, in Pakistan and Nigeria, progress is slower. In Pakistan, although the higher courts have begun implementing case management tools and videoconferencing, cultural barriers persist, along with unstable rural internet infrastructure and a lack of specific legal frameworks for ODR (13). In Nigeria, infrastructure remains predominantly analog, and while the 2023 Arbitration and Mediation Act modernized the general framework, it still lacks explicit provisions governing the enforcement of awards obtained entirely through digital processes (14).

In conclusion, the global panorama shows that while some jurisdictions are building "digitally native justice" ecosystems, others are still struggling to close the digital literacy gap and overcome basic infrastructure shortfalls (19, 13).

III. Soft Law as a Regulatory Response

The speed with which artificial intelligence has been integrated into legal practice has outpaced the capacity of national legislators to respond, creating a regulatory vacuum that international institutions have sought to fill through soft law (2, 8). These non-binding instruments serve as essential reference frameworks for harmonizing standards of conduct, ensuring that the use of technology does not compromise the integrity of the process or the validity of the resulting awards or agreements (4, 5). The transnational nature of arbitration and mediation makes these guidelines essential to providing legal certainty in the face of state regulations that may vary significantly or simply not exist (5, 8).

3.1. The SVAMC Guidelines (2024): the pioneering precedent

The Silicon Valley Arbitration and Mediation Center (SVAMC) marked a milestone in April 2024 by publishing its Guidelines on the Use of Artificial Intelligence in Arbitration, the first comprehensive, principles-based framework for the industry (5, 17). These guidelines do not seek to replace local laws, but rather to serve as a supplementary international standard promoting fairness, efficiency, and transparency (5).

The SVAMC guidelines are structured around three pillars:

• Guidelines for all participants: they impose a responsibility to become familiar with the uses, limitations, and risks of AI applications, including bias and the "black box problem" (5). They underscore the duty to safeguard confidentiality, prohibiting the entry of sensitive data into tools that have not been properly audited and authorized (5).

• Guidelines for parties and their representatives: they establish a duty of competence and diligence, under which lawyers bear ultimate responsibility for any error or "hallucination" generated by AI in their submissions (5). They likewise prohibit the use of AI to fabricate evidence or distort the authenticity of facts (5).

• Guidelines for arbitrators: this is perhaps the most critical point, stipulating the non-delegation of decision-making responsibilities (5). An arbitrator may use AI to organize information but may not substitute a machine's output for their own independent analysis of the facts and the law (5).

3.2. The CIArb Guideline (2025): the global standard for arbitration

In 2025, the Chartered Institute of Arbitrators (CIArb) published its Guideline on the Use of AI in Arbitration, consolidating a professional standard of global reach (4). Unlike other documents, this guideline goes deeper into critical technical definitions such as machine learning and natural language processing, providing practitioners with a solid terminological foundation (4).

The CIArb guideline introduces the concept of "high-risk AI use," defined as use that has the potential to exert non-human influence over the award or to violate the parties' privacy (4). Among its most notable contributions are:

• Transparency and disclosure: it establishes that high-risk AI use must be disclosed in writing to the opposing party and the tribunal, detailing the tool used and the parameters of the prompt or instruction (4).

• Tribunal powers: it expressly recognizes arbitrators' authority to issue procedural orders regulating the use of technology, appoint independent AI experts, and, in case of non-compliance, draw adverse inferences or adjust cost awards (4).

• Ownership of the decision: it reaffirms that the arbitrator bears full responsibility for every aspect of the award, regardless of any technological support received during the analysis or drafting phase (4, 20).

3.3. The IBA Mediation Guidelines (2025): closing the gap

While arbitration has focused on adjudication, mediation requires a different approach grounded in facilitating agreement (6, 7). The International Bar Association (IBA) Mediation Committee responded to this need in 2025 with its Guidelines on the use of generative AI in mediation (6).

These guidelines focus specifically on generative AI, recognizing its potential to transform mediation through rapid document synthesis, the generation of resolution options, and real-time language support (6). However, they establish strict safeguards:

• Neutrality and impartiality: mediators must ensure that AI use does not introduce biases favoring one party or compromising their independence (6). It is recommended to vary prompt wording and compare outputs across different models to mitigate this risk (6).

• Party autonomy: since mediation is consensual, the parties retain full control over the process. The guidelines stipulate that any participant may request information about the tools others intend to use, and may demand their discontinuation if there are well-founded doubts about their integrity (6).

• Data protection and agreements: when using AI to draft settlement agreements, the guidelines require mandatory human review by legal professionals to ensure the text faithfully reflects the parties' intent and is enforceable under frameworks such as the 2019 Singapore Convention on International Settlement Agreements Resulting from Mediation (6, 14).

Taken together, this soft law ecosystem has succeeded in creating a procedural safety net that allows the benefits of AI —such as cost reductions of 30–50% (20)— to be captured without sacrificing the fundamental principles of human justice (4, 5, 6).

IV. Jurisdictional Analysis and Implementation Models

The integration of artificial intelligence (AI) into alternative dispute resolution (ADR) is not a monolithic process; rather, it manifests through diverse models reflecting each jurisdiction's policy priorities and technological maturity (9, 11). While some nations have adopted a fully state-integrated approach, others have opted for more cautious governance frameworks or relied on private-sector dynamism to fill gaps left by traditional judicial systems (9, 16).

4.1. The Asian model: China and Singapore

Asia has positioned itself as the epicenter of digital justice, offering two contrasting yet equally influential models for the institutional design of dispute resolution (9).

• China: the "Smart Courts" and full-integration model: China has implemented the most ambitious judicial modernization project globally, where Online Dispute Resolution (ODR) is not an ancillary mechanism but a core component of judicial governance (9). Internet courts (Hangzhou, Beijing, and Guangzhou) manage the entire litigation cycle —from blockchain-based evidence submission to judgment-drafting assistance— under a scheme of massive procedural standardization (9). This model stands out for its capacity to process high volumes of e-commerce and intellectual property disputes, although it raises critical questions about whether systemic efficiency is taking precedence over party autonomy (9).

• Singapore: supervised modularity and regulatory rigor: unlike China's centralized approach, Singapore has adopted "supervised modularity" (9). Through the Community Justice and Tribunals System (CJTS), the country facilitates digital access for filing claims and offers e-Negotiation and eMediation processes (9). However, Singapore maintains a strict distinction between technological facilitation and adjudication, prioritizing soft law frameworks that ensure generative AI use by litigants does not displace fundamental legal and professional obligations (9).

4.2. Adoption leaders: UAE and India

• United Arab Emirates (UAE): the UAE stands at the global forefront of AI adoption in the workplace, with a generative AI tool usage rate that reached 70.1% in the first quarter of 2026 (19). This leadership is the result of a state strategy launched in 2017 that has integrated AI into legislation and public administration (19). In the arbitral sphere, institutions in Dubai and Abu Dhabi have modernized their rules to encourage technology use and improve case management efficiency (2).

• India: innovation in the face of judicial backlog: with a backlog of more than 50 million pending cases in its courts, India has found in AI and ODR a necessary lever for transformation (16, 20). Private platforms such as Presolv360, backed by recommendations from institutions like NITI Aayog, have resolved more than a million disputes with time and cost savings of up to 65% compared to traditional litigation (16). Implementation in India is characterized by a vibrant legaltech ecosystem seeking to democratize access to justice in a high-volume market (16, 20).

4.3. Perspectives from Latin America: the case of Chile

Chile has established itself as the undisputed regional leader, topping the Latin American Artificial Intelligence Index (ILIA) (11). With a projected sector growth rate of 33% annually, the country has developed pioneering tools both in the judicial and arbitral spheres (11).

The "Justa" project stands out in family and labor courts, using algorithms for transcription and document analysis, drastically reducing procedural timelines (11). In the private sector, CAM Santiago has implemented "Online Resolution" systems that enable algorithm-assisted negotiations for consumer disputes (11). Chile is also advancing legislative proposals on AI that follow the European Union's risk-based model, which could potentially classify dispute resolution systems as "high-risk" in order to ensure transparency and human oversight (11).

4.4. Emerging markets: Pakistan and Nigeria

In developing jurisdictions, AI is seen as an opportunity to overcome infrastructural gaps, although significant structural challenges remain (13, 14).

• Pakistan: the higher courts of Islamabad and Lahore have implemented case management tools and e-filing systems that have improved operational transparency (13). However, widespread adoption is limited by the digital divide in rural areas and a strong cultural preference for in-person dispute resolution (13). Pakistan still lacks a specific legal framework for ODR, creating uncertainty over the validity of certain purely digital processes (13).

• Nigeria: the 2023 Arbitration and Mediation Act (AMA) has modernized the Nigerian legal framework, allowing the use of electronic communications and digital signatures (14). Nevertheless, Nigeria faces challenges related to unstable internet infrastructure and the need for greater training of legal professionals in the use of technological tools (14). As in Pakistan, the success of the Nigerian model will depend on establishing clear standards for enforcing awards obtained on fully digital platforms (14).

V. Tools and Platforms: Technology in Action

The transition toward advanced digital justice has ceased to be a project of administrative modernization and has become an operational ecosystem in which artificial intelligence (AI) takes on critical functional roles (9). Today, technological tools do not merely facilitate communication; they actively intervene in case triage, evidence management, settlement-proposal design, and adjudicative decision support (9). This section examines the materialization of these technologies through automated platforms, institutional offerings from leading bodies, and the economic dynamism underpinning this innovation.

5.1. Automated arbitration platforms

The deployment of automated arbitration platforms represents the most disruptive advance in the architecture of ADR (11). A milestone in this evolution is the launch, by the American Arbitration Association (AAA) and its international division (ICDR), of an AI-powered arbitrator in November 2025 (11). Unlike traditional assistance tools, this system is capable of evaluating the merits of a case, generating recommendations, and preparing draft awards (11).

This model, initially implemented for document-only construction disputes, operates under a "power-assisted co-pilot" scheme (13). The system has been trained on a dataset of more than 1,500 annotated construction awards, ensuring high quality in reasoning and contextual precision (13). Nevertheless, the design preserves human primacy: human arbitrators supervise, validate, and sign off on final decisions, while also allowing the parties to confirm the AI's understanding of their submissions before the draft is produced (11, 13).

In parallel, private platforms such as Presolv360 in India have demonstrated the scalability of these models (16). This platform integrates AI workflows for negotiation, mediation, and arbitration through an interface designed to process high volumes of disputes, with cost and time savings of up to 65% compared to traditional litigation (16, 20). In the public sphere, China's internet courts have normalized the use of tools that automate the entire judicial cycle, including identity verification and blockchain-based evidence preservation (9).

5.2. Institutional tools: the AAA case

The American Arbitration Association (AAA) has positioned itself as the global benchmark for institutional AI integration, adopting an "AI-native" vision for dispute resolution (11). Beyond adjudication, the AAA has focused its technological efforts on improving transparency and precision in the selection of neutral third parties (3).

In October 2025, the institution launched AAAi Mediator Search, a digital platform that uses AI to identify mediators based on highly specific criteria, such as subject-matter experience, geographic location, language proficiency, and fee rates (3, 4). This tool ranks results by relevance and allows parties to access profiles and résumés directly, simplifying a process that traditionally required exhaustive manual research (4).

This offering is complemented by AAAi Panellist Search, introduced in 2024, which assists case managers in identifying arbitrator candidates during the nomination process (12). These tools reflect an institutional strategy aimed at providing greater clarity, choice, and confidence to parties involved in mediation and arbitration proceedings (5).

5.3. LegalTech investment and market

Technological growth in ADR is underpinned by vigorous economic expansion. The global dispute resolution solutions market, valued at $11.1 billion in 2025, is projected to reach $19.8 billion by 2034, with a compound annual growth rate (CAGR) of 6.7% (3). This growth is driven by the increasing complexity of global commercial relationships and the high cost of traditional litigation (3).

The flow of capital into the LegalTech sector has reached record levels. In 2024, global venture capital investment in these technologies rose to $2.2 billion, with AI-based solutions capturing the bulk of this funding (16). This dynamism is especially notable in markets such as India, where platforms like Presolv360 have secured Series A funding rounds of $4.7 million to scale their AI-driven workflows and expand enterprise adoption (16).

At the end-user level, adoption of generative AI tools is gaining ground rapidly (19). In early 2026, global usage of these tools among the working population reached 17.8%, with peaks of up to 70.1% in leading jurisdictions such as the United Arab Emirates (19). This trend suggests that the market demands not only efficiency but full AI integration enabling disputes to be resolved quickly, fairly, and cost-effectively in an interconnected business environment (3, 19).

VI. Ethical and Legal Challenges

The integration of artificial intelligence (AI) into arbitration and mediation is not merely a technical advance, but a structural challenge to the foundations of private justice (7, 9). As AI tools move from being support instruments to participating in the formation of normative judgments, critical tensions emerge that test the principles of fairness, neutrality, and due process (9, 20). The international legal community warns that the success of AI should not be measured by efficiency alone, but by its ability to withstand a "procedural legitimacy test" (9).

6.1. Algorithmic bias and due process

One of the most insidious risks of AI is the perpetuation and amplification of pre-existing biases in training datasets (4, 5). AI tools do not possess intelligible reasoning; they produce outputs based on complex probabilistic calculations that can replicate patterns of historical discrimination (5, 9). This "algorithmic bias" is particularly sensitive in the selection and appointment of arbitrators or experts, where the underrepresentation of certain groups can be exacerbated by algorithms that favor traditional profiles based on past data (5).

From a due process perspective, AI use raises the challenge of "ownership of the decision" (4, 20). There is a risk that human arbitrators, drawn by efficiency, may inadvertently delegate their personal mandate to the machine, falling into "cognitive inertia" or confirmation bias regarding algorithm-generated outputs (4, 15). International guidelines are unequivocal: the decision-making function cannot be delegated (5). An award based on a factual analysis distorted by an AI "hallucination" —where the system invents precedents or facts to fill information gaps— violates the parties' right to a fair and well-founded determination (4, 17). Moreover, the insertion of AI into case triage may "narrow" the dispute too early, preventing a party from properly presenting its theory of the case if it does not fit the categories predefined by the system (9).

6.2. Confidentiality and data protection

Confidentiality is the cornerstone of arbitration and mediation, and preserving it in the AI era is extremely complex (5, 8). Entering sensitive information, trade secrets, or personal data into third-party AI tools —particularly open-source large language models (LLMs)— carries the risk that such information may later be used to train the model or become accessible to external users in the future (5, 6).

Data protection regulations, such as China's PIPL, Singapore's PDPA, or Chile's new legislation, impose strict obligations regarding data security and minimization (9, 11). Participants in dispute resolution proceedings must conduct due diligence on the data retention policies of the platforms used, opting for closed, secure solutions (5). In jurisdictions with developing digital infrastructure, such as Nigeria, the lack of minimum technical standards heightens vulnerability to cyberattacks and data breaches, which not only compromises privacy but can undermine the integrity of the entire arbitral proceeding (14, 20). Drafting mediation agreements with AI therefore requires a process of "anonymization" and rigorous human review to prevent protected data from leaking into public databases (6).

6.3. Transparency and accountability in AI use

The "black box" problem constitutes the greatest obstacle to transparency: the inability of users —and even developers— to understand the exact mechanism by which an AI reaches a specific outcome (4, 5, 9). This opacity collides with arbitrators' duty to provide clear and reasoned justification for their decisions (5). Without "explainable AI," parties may be forced to accept recommendations or decisions whose internal logic is inaccessible, weakening trust in the justice system (5, 9).

In response, current regulatory frameworks propose a "continuous disclosure" standard (4, 5, 6). Parties must be informed when high-risk AI is used, defined as AI that exerts a significant influence on the process or outcomes (4). Responsibility, however, always remains with the human actor (4, 5, 20). Lawyers are responsible for the accuracy of AI-generated submissions, and arbitrators must take full authorship of the award, regardless of the technological support received (4, 5). Ultimately, the legitimacy of ADR in this new environment will depend on technology acting as a "power-assisted co-pilot" under strict human supervision, ensuring that the pursuit of speed does not diminish the dignity and autonomy of the parties involved (9, 20).

VII. Conclusions and Future Perspectives

The integration of artificial intelligence (AI) into international mediation and arbitration has ceased to be a technological promise and has become a functional reality reshaping the architecture of private justice (9). Following the comprehensive analysis conducted in this article, the following key conclusions and strategic projections can be drawn regarding the future of alternative dispute resolution (ADR).

Synthesis of findings: from efficiency to legitimacy

• Structural transformation of the market: the evolution of ADR is intrinsically tied to the expansion of the technology solutions market, projected to reach $19.8 billion by 2034 (3). This expansion is not merely quantitative but qualitative: AI has moved from simply assisting communication to systems participating in triage, evidence management, and adjudicative decision support (9).

• The triumph of soft law as a governance framework: in the absence of binding state regulations, instruments such as the SVAMC Guidelines (2024), the CIArb Guideline (2025), and the IBA Guidelines (2025) have succeeded in establishing a global due diligence standard (4, 5, 6). The guiding principle is clear: technology can be a "power-assisted co-pilot," but ethical judgment and final responsibility for the award or settlement agreement remain non-delegable for the neutral third party (11, 20).

• Jurisdictional divergence and the digital divide: the global panorama reveals asymmetric development. While the United Arab Emirates (19) and Singapore (9) consolidate native digital justice ecosystems, and Chile leads the Latin American region with pioneering judicial AI projects (11), emerging markets such as India face the challenge of scaling these tools to resolve massive case backlogs (16, 20). A worrying gap nonetheless persists between the Global North and the Global South, driven by disparities in infrastructure and digital literacy (13, 19).

• Ethical challenges as enforcement barriers: algorithmic opacity —the "black box problem"— and biases in training data represent the greatest risks to due process (4, 5, 9). The validity and enforceability of awards under frameworks such as the 1958 New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards will critically depend on AI being used with full transparency and under effective human oversight that prevents the inadvertent delegation of the decision-making mandate (12, 17, 20).

Future perspectives: toward a disciplined digital justice system

The future of arbitration and mediation is not heading toward a "court of machines," but toward a "procedurally disciplined digital justice system" (9). In this scenario, three critical trends are anticipated for the coming years:

First, the consolidation of advanced institutional tools. Organizations such as the AAA-ICDR have already paved the way with the launch of AI-powered mediators and arbitrators (2, 11). By 2026, these tools are expected to expand into higher-value cases and more complex sectors, enabling cost reductions of between 30% and 50% (20).

Second, the evolution of the "procedural legitimacy" test. The success of AI will no longer be measured solely by speed or savings, but by its capacity to be auditable and explainable (9). The industry will need to move toward creating "automation logs" detailing AI functions and the data parameters used, ensuring that party autonomy is not eroded by invisible algorithmic nudges (9).

Finally, the need for robust international harmonization. To avoid a patchwork of conflicting regulations that would jeopardize legal certainty, the international community must work toward cross-border interoperability standards for ODR (9). Only then will AI be able to fulfill its promise of democratizing access to justice and resolving disputes in real time within an interconnected global economy (3, 16).

In conclusion, artificial intelligence is rewriting the rules of ADR (18). The challenge for legal professionals and policymakers is not to resist automation, but to ensure that the pursuit of technological efficiency does not come at the expense of the integrity and fundamental rights of those seeking justice (9, 20).

Bibliography and Consolidated Reference List

In accordance with the mandatory citation system defined for this research, the complete list of sources used in drafting this article is presented below:

(1) Queen Mary University of London / White & Case. 2025 International Arbitration Survey: The path forward: Realities and opportunities in arbitration. 2025.

(2) Freshfields. Shifting terrain: the trends transforming international arbitration. 2026.

(3) Research and Markets. Conflict Resolution Solutions Market Outlook 2025-2034: Market Share, and Growth Analysis. 2025.

(4) Chartered Institute of Arbitrators (CIArb). Guideline on the Use of AI in Arbitration. 2025.

(5) Silicon Valley Arbitration and Mediation Center (SVAMC). Guidelines on the Use of Artificial Intelligence in Arbitration. 2024.

(6) International Bar Association (IBA) Mediation Committee. Guidelines on the use of generative AI in mediation. 2025.

(7) Rodríguez-Salcedo et al. "Rethinking Mediation and Arbitration in the Age of Artificial Intelligence: A Systematic Review". Dialnet, 2025.

(8) "La inteligencia artificial en el arbitraje internacional 2.0. Oportunidades y desafíos en un futuro que ya es presente" [Artificial intelligence in international arbitration 2.0: opportunities and challenges in a future that has already arrived]. Cuadernos de Derecho Transnacional, UC3M, 2025.

(9) "AI-Enabled Online Dispute Resolution in China and Singapore: Institutional Design, Procedural Safeguards and Comparative Lessons". International Journal of Law Management & Humanities, 2026.

(10) Thomson Reuters / Westlaw. "Artificial Intelligence in Dispute Resolution: Developments, Challenges and Perspectives for Legal Practice". Secondary Sources, 2025.

(11) Clyde & Co. "AI in Arbitration: A Perspective from Chile". 2025.

(12) CMS.law. "AI in arbitration and the courts: Focus on Türkiye & Ukraine". 2025.

(13) "Exploring the Efficacy of Online Dispute Resolution Mechanisms: A Case Study of Pakistan". 2025.

(14) "Online Dispute Resolution (ODR) and the Future of Digital Commercial Arbitration and Mediation in Nigeria". Journals Place, 2025.

(15) "We just built Corporate America's dream and Asimov's nightmare—a machine that automates all dispute resolution. Meet Arbit". Mediate.com, 2025.

(16) "Presolv360 Secures $4.7 Million to Scale India's Digital Dispute Resolution Market". 2025.

(17) American Arbitration Association (AAA). "AAA® Launches AI-Powered Mediator Search Tool". 2025.

(18) American Arbitration Association (AAA). "The Dispute Revolution: How AI is Rewriting the Rules of ADR". 2025.

(19) "UAE leads the world in AI adoption with 70% workplace usage". Middle East AI News, 2025.

(20) "AI Arbitrators: Ushering in a New Era of International Arbitration". India Law Chronicle, 2025.