Civil Law and AI Arbitration Platforms Regulation

1. Introduction

Artificial intelligence (AI) arbitration platforms are digital systems that assist parties in resolving civil and commercial disputes through automated negotiation, online dispute resolution (ODR), algorithmic analysis, and, in some cases, automated decision-making.

Traditional arbitration involves an independent arbitrator who considers the evidence and arguments of the parties and issues an award. An AI-enabled arbitration platform may perform some or all of the supporting functions, including document analysis, legal research, procedural scheduling, evidence classification, prediction of likely outcomes, and drafting proposed awards.

The development of these platforms raises important questions of civil law:

Can an AI system lawfully perform the functions of an arbitrator?

Who is liable if an algorithm produces an incorrect decision?

How can parties challenge an award affected by AI bias or errors?

What duties do platform operators owe to parties?

How should confidentiality, personal data, cybersecurity, and due process be protected?

Can an AI-generated decision be recognized and enforced as an arbitral award?

The central legal principle is that using AI in arbitration does not eliminate the requirements of consent, impartiality, procedural fairness, and judicial supervision. Whether an AI-assisted or AI-generated decision qualifies as a legally enforceable award depends on the applicable arbitration law, the parties' agreement, and the circumstances of the proceedings.

This explanation examines the subject principally through comparative civil and commercial arbitration law, with examples from the United States, England, France, Singapore, and other jurisdictions. The case laws below are real judicial decisions relevant to AI arbitration regulation, although most predate modern AI arbitration platforms and therefore provide analogous legal principles rather than direct rulings on AI arbitrators.

2. Meaning and types of AI arbitration platforms

AI arbitration platforms are technological systems used to facilitate or conduct dispute resolution between parties who have agreed to arbitrate.

A. AI-assisted arbitration

AI organizes evidence, summarizes pleadings, identifies legal authorities, translates documents, and helps a human arbitrator manage the proceedings. The arbitrator retains responsibility for findings and the final award.

B. Online arbitration platforms

These platforms facilitate digital filing, virtual hearings, electronic evidence submission, notices, scheduling, and delivery of awards. AI may be an optional feature rather than the decision-maker.

C. Algorithmic decision-support systems

These systems evaluate claims, compare evidence, estimate damages, or predict likely outcomes. Their predictions must not be confused with legally binding arbitral findings.

D. Automated or AI-led arbitration

The system may generate a proposed resolution or draft award. The legally significant question is whether the applicable law and arbitration agreement permit the process, and whether a legally qualified and properly appointed decision-maker must make the final determination.

A useful distinction is that an online arbitration platform is not necessarily an AI arbitrator. A platform can digitize the entire procedure while leaving adjudication entirely to a human.

3. Legal framework governing AI arbitration platforms

A. The arbitration agreement and party consent

Arbitration is generally founded on the parties' agreement to submit specified disputes to arbitration. The agreement may identify an arbitral institution, appoint an arbitrator, establish procedural rules, and specify the seat of arbitration.

AI raises an additional question: what exactly did the parties consent to?

For example, an agreement providing that disputes will be resolved by a named human arbitrator using an online platform does not automatically authorize an AI system to replace that arbitrator. Similarly, an agreement to use an AI-assisted process does not necessarily mean that a party has consented to a final decision produced without meaningful human adjudication.

Civil courts may examine:

Whether the arbitration agreement is valid and sufficiently clear.

Whether the dispute falls within its scope.

Whether the tribunal was constituted in accordance with the agreement.

Whether an AI system merely assisted the tribunal or effectively replaced it.

Whether the procedure complied with mandatory law and applicable institutional rules.

B. The UNCITRAL Model Law

The UNCITRAL Model Law on International Commercial Arbitration provides a significant framework for many jurisdictions.

Its provisions are especially relevant to AI arbitration:

The Model Law does not itself constitute a comprehensive AI-arbitration code. Its existing safeguards can nevertheless apply when AI affects tribunal composition, procedure, evidence, reasoning, or the validity of an award.

For example, if an AI tool generates a factual summary that materially influences an award, the tribunal should ensure that the parties have a fair opportunity to challenge disputed factual assumptions. An algorithm cannot legitimately be used to introduce decisive evidence that neither party has had an opportunity to address.

C. The New York Convention, 1958

The Convention on the Recognition and Enforcement of Foreign Arbitral Awards is central to international arbitration.

Two provisions are particularly important:

Article V(1)(d): Recognition or enforcement may be refused where the composition of the arbitral authority or the arbitral procedure did not accord with the parties' agreement or, failing such agreement, the law of the country where arbitration took place.

Article V(1)(b): Enforcement may be refused where a party was unable to present its case.

These provisions may become relevant if a purported AI arbitrator was not appointed in accordance with the agreement or if undisclosed algorithmic reasoning deprived a party of a fair opportunity to contest the case.

However, an AI-related irregularity does not automatically invalidate an award. The court must assess the applicable statutory grounds and the facts of the dispute.

D. Institutional arbitration rules

Arbitration institutions establish rules concerning tribunal appointment, disclosure, procedure, evidence, and awards. AI use must be assessed against the rules selected by the parties, as well as any mandatory legal requirements.

A significant recent development is the American Arbitration Association's Digital Dispute Resolution Center and its AI-led arbitration rules. Its published terms describe AI-assisted summaries, timelines, case analyses, and draft awards, with human review and the human arbitrator issuing the final award in binding cases. This illustrates a regulatory approach that permits substantial automation while retaining human responsibility.

American Arbitration Association

The Chartered Institute of Arbitrators' 2025 Guideline on the Use of AI in Arbitration also addresses disclosure, bias, confidentiality, data quality, and procedural fairness.

Jus Mundi

These instruments are not universal statutes. Their applicability depends on the institution, contractual adoption, and relevant law.

4. Core civil-law and regulatory issues

A. Independence and impartiality

An arbitrator must be independent and impartial under the applicable legal framework. AI creates additional risks when:

The system was developed or supplied by a party to the dispute.

Its training data systematically favor a particular type of claimant.

The platform operator has a financial interest in a particular outcome.

A party has privileged access to the algorithm or its underlying information.

The arbitrator relies on an AI-generated recommendation without adequately examining it.

A system does not become impartial merely because it uses mathematical calculations. Bias can arise from training data, model design, data selection, or the way a question is framed.

Regulatory safeguard: Require appropriate disclosure of conflicts of interest, material limitations, and relevant algorithmic dependencies. Where the circumstances justify it, an independent audit may be necessary.

B. Procedural fairness and natural justice

The principles of audi alteram partem (hear the other side) and nemo judex in causa sua (no one should judge their own cause) are relevant to AI-assisted arbitration.

Suppose an AI system classifies a company's conduct as fraudulent because its model identifies a pattern in transaction data. If the tribunal relies on that classification without examining the underlying transactions or allowing the company to respond, the process may be procedurally unfair.

The central questions are:

Was the evidence disclosed or otherwise fairly available?

Could the parties challenge the AI's assumptions and conclusions?

Did the tribunal independently evaluate the arguments?

Were the findings based on the evidence and applicable law?

Was a reasoned and legally valid award issued?

A hearing need not always be oral; the applicable law and agreed procedure determine the appropriate form. What matters is that each party receives a meaningful opportunity to present its case.

C. Transparency and explainability

Some AI models cannot provide a complete or reliable account of how they reached a particular result. This creates a problem when the system materially affects the tribunal's findings.

A platform should, where appropriate, preserve:

The version of the model used.

Relevant prompts, inputs, and outputs.

The sources supporting generated legal propositions.

The evidentiary material underlying material factual summaries.

Records of human verification and corrections.

Transparency does not necessarily mean that all proprietary source code must be disclosed. The level of disclosure should be proportionate to the significance of the AI's role, the need to protect confidential information, and the parties' ability to challenge the result.

D. Data protection and confidentiality

Arbitration frequently involves confidential contracts, trade secrets, financial records, personal information, and commercially sensitive correspondence.

Uploading such information to an external AI platform may create risks involving unauthorized access, retention, model training, or cross-border data transfers.

Platform operators and participants should therefore consider:

Data minimization and purpose limitation.

Encryption and access controls.

Contractual restrictions on reuse of case materials.

Retention and deletion policies.

Applicable data-protection legislation.

Incident reporting and breach response.

Confidentiality is not an absolute exemption from privacy or data-protection law. The applicable statutory obligations must still be assessed.

E. Liability for AI errors

AI platforms may generate fabricated legal authorities, misstate contract clauses, overlook evidence, or calculate damages incorrectly.

Civil liability depends on the relevant contract, applicable law, the duty breached, causation, and the recoverable loss. Depending on the circumstances, potential defendants may include the platform operator, software developer, arbitrator, professional adviser, or contracting party.

For example, if a platform incorrectly calculates interest and the arbitrator adopts the calculation without verification, the legal analysis should distinguish among:

The software defect.

The operator's contractual obligations.

The arbitrator's procedural and adjudicative responsibilities.

The effect of the error on the award.

Whether the loss is legally recoverable from the particular defendant.

The mere presence of an AI error does not automatically establish negligence or create liability for every participant.

F. Recognition, enforcement, and challenges to awards

An award affected by AI may be challenged under existing arbitration law if the facts establish a recognized ground for setting it aside or refusing enforcement.

Possible issues include improper tribunal composition, excess of authority, serious procedural irregularity, inability to present a case, or a conflict with applicable public policy.

The distinction is important: a court ordinarily does not set aside an award merely because the arbitrator used AI. The relevant question is whether the use of AI resulted in a legally material defect under the governing law.

5. At least 8 important case laws

The following cases provide judicial principles relevant to the regulation of AI arbitration platforms. They are not presented as cases that directly decided the legality of an AI arbitrator. Instead, they establish legal rules concerning consent, independence, procedural fairness, tribunal authority, and judicial review that can be applied to AI-enabled arbitration.

Case 1. First Options of Chicago, Inc. v. Kaplan (1995)

United States Supreme Court · 514 U.S. 938 (1995)

Facts: A dispute arose over whether certain parties were bound by an arbitration agreement. The parties disputed who should decide whether the arbitration issue itself had been submitted to arbitration.

Legal issue: When should a court defer to an arbitrator's determination of arbitrability?

Judgment and principle: The US Supreme Court held that courts generally determine arbitrability independently unless the parties clearly and unmistakably agreed to submit that question to the arbitrator.

Relevance to AI arbitration platforms: This case supports the importance of clear consent. A platform's standard terms should not be treated as proof that the parties agreed to autonomous AI adjudication. If a party disputes whether the agreement authorized AI to make a binding decision, the question must be resolved under the governing arbitration law and applicable agreement.

Civil-law significance: Consent determines the scope of arbitral authority. Technology cannot expand that authority simply because the platform's software is capable of doing more than the parties agreed.

Case 2. Commonwealth Coatings Corp. v. Continental Casualty Co. (1968)

United States Supreme Court · 393 U.S. 145 (1968)

Facts: An arbitration award was challenged because an arbitrator had a business relationship with one of the parties that had not been disclosed.

Legal issue: What degree of disclosure is required to protect impartiality in arbitration?

Judgment and principle: The Court set aside the award, emphasizing the importance of impartiality and disclosure of relationships that may create justifiable doubts about an arbitrator's neutrality.

Relevance to AI arbitration platforms: An analogous problem can arise if a platform's developer has a commercial relationship with one party, if the system was specifically customized for that party, or if the algorithm's design creates a material conflict of interest.

The case does not establish that every software-provider relationship must be disclosed. Its relevance lies in the broader principle that circumstances capable of undermining confidence in impartial adjudication must be examined.

Civil-law significance: Disclosure and impartiality are safeguards of legitimate dispute resolution, not merely technical formalities.

Case 3. Hall Street Associates, L.L.C. v. Mattel, Inc. (2008)

United States Supreme Court · 552 U.S. 576 (2008)

Facts: The parties agreed that a court could review an arbitration award for legal error, going beyond the review grounds provided by the Federal Arbitration Act.

Legal issue: Can parties expand the statutory grounds on which a federal court may vacate an arbitration award?

Judgment and principle: The Supreme Court held that the Federal Arbitration Act's statutory grounds for vacatur and modification were exclusive in the circumstances before it. Parties could not use their agreement to create the broader judicial review standard at issue.

Relevance to AI arbitration platforms: Parties should not assume that an AI-related error automatically permits a full rehearing on the merits. A challenge must fit the applicable statutory framework.

For example, if an AI-generated draft contains an incorrect legal proposition, a court's power to set aside the resulting award will depend on the applicable law and the nature of the defect—not simply on the fact that AI was involved.

Civil-law significance: The regulation of AI arbitration must account for the limited scope of post-award judicial review and the importance of identifying procedural defects before the award becomes final.

Case 4. Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc. (1985)

United States Supreme Court · 473 U.S. 614 (1985)

Facts: A dispute arising from an international commercial relationship raised questions about arbitrating statutory antitrust claims.

Legal issue: Can international arbitration be used to resolve claims arising under statutory law?

Judgment and principle: The Court enforced the agreement to arbitrate the statutory claims in the case, recognizing the capacity of international arbitration to address such disputes.

Relevance to AI arbitration platforms: The case illustrates that arbitration can operate within a broader statutory legal framework. AI-assisted arbitration is therefore not exempt from mandatory laws merely because parties choose a technological dispute-resolution mechanism.

For instance, an AI platform handling a consumer or competition dispute must be evaluated against applicable statutory protections and restrictions on arbitrability.

Civil-law significance: Party autonomy does not automatically displace substantive statutory obligations. An AI platform must operate within the legal framework governing the underlying dispute.

Case 5. Rent-A-Center, West, Inc. v. Jackson (2010)

United States Supreme Court · 561 U.S. 63 (2010)

Facts: An employee challenged an arbitration agreement that delegated questions of arbitrability to the arbitrator. The dispute concerned the allocation of authority to decide challenges to the arbitration agreement.

Legal issue: When can a delegation clause require the arbitrator, rather than the court, to decide certain threshold issues?

Judgment and principle: The Court enforced the delegation provision because the challenge before it was directed at the arbitration agreement generally rather than specifically at the delegation clause, subject to the applicable rules of arbitration law.

Relevance to AI arbitration platforms: AI platform agreements may contain provisions assigning particular procedural questions to an arbitral tribunal or defining the scope of a dispute-resolution process. Their enforceability cannot be presumed merely because they appear in digital terms of service.

A court must consider the governing law, the wording of the agreement, and the nature of the challenge. In jurisdictions with different statutory rules, the outcome may differ.

Civil-law significance: The allocation of decision-making authority is a matter of agreement and law, not a feature that a software provider may unilaterally determine.

Case 6. TRF Ltd. v. Energo Engineering Projects Ltd. (2017)

Supreme Court of India · (2017) 8 SCC 377

Facts: An arbitration clause allowed a person who was ineligible to act as arbitrator to nominate another arbitrator.

Legal issue: Could a person who was legally ineligible to act as an arbitrator retain the power to nominate the arbitrator?

Judgment and principle: The Supreme Court of India held that a person rendered ineligible to act as arbitrator under the applicable statutory framework could not nominate another arbitrator.

Relevance to AI arbitration platforms: The decision is relevant to the integrity of tribunal appointment. An AI system cannot independently resolve the question of its own legal eligibility. Where the governing law requires a qualified and eligible arbitrator, a platform cannot bypass those requirements through automated appointment or delegation mechanisms.

The case does not directly decide whether an AI can ever qualify as an arbitrator; that question requires separate statutory analysis.

Civil-law significance: The appointment mechanism must comply with mandatory rules governing eligibility and independence.

Case 7. Perkins Eastman Architects DPC v. HSCC (India) Ltd. (2019)

Supreme Court of India · (2020) 20 SCC 760

Facts: A dispute concerned an arbitration clause that gave one party's designated official the power to appoint the sole arbitrator.

Legal issue: Could a person interested in the outcome of the dispute exercise exclusive control over the appointment of the arbitrator?

Judgment and principle: The Supreme Court held that a person interested in the outcome of the dispute could not have exclusive authority to appoint the sole arbitrator in the circumstances before it.

Relevance to AI arbitration platforms: Consider a commercial platform that allows one contracting party to select the AI model, determine its parameters, and control the appointment process. That arrangement may raise concerns about equality and impartiality, depending on the governing law and facts.

The case supports scrutiny of appointment structures that place one interested party in a position of disproportionate control.

Civil-law significance: Neutrality must be protected at the point of appointment as well as during the proceedings.

Case 8. Ssangyong Engineering & Construction Co. Ltd. v. National Highways Authority of India (2019)

Supreme Court of India · (2019) 15 SCC 131

Facts: The dispute concerned a construction contract and an arbitral award that was challenged under Section 34 of the Arbitration and Conciliation Act, 1996.

Legal issue: How should courts apply the statutory grounds for setting aside arbitral awards, including the public-policy standard and fundamental procedural requirements?

Judgment and principle: The Supreme Court examined the limits of judicial intervention under the amended Arbitration and Conciliation Act and addressed serious defects in the award and its treatment of the contractual dispute.

Relevance to AI arbitration platforms: An award that relies on AI-generated evidence or reasoning may face a challenge if the circumstances establish a recognized statutory ground, such as a fundamental procedural defect or a conflict with public policy within the applicable legal standard.

The decision does not permit courts to set aside awards merely because they disagree with an arbitrator's interpretation or because AI was used.

Civil-law significance: AI must not be used to circumvent the limited but important safeguards governing arbitral awards under Indian law.

6. What the eight cases establish collectively

These cases are particularly useful because they show that the principal legal safeguards already exist in arbitration law. The difficult task is applying them to new technologies without assuming either that AI is inherently unlawful or that automation removes the need for procedural protections.

7. Indian civil law and AI arbitration platform regulation

India provides a particularly useful example because the Arbitration and Conciliation Act, 1996 already contains provisions addressing tribunal appointment, procedural fairness, arbitral awards, and judicial challenges.

A. Arbitration and Conciliation Act, 1996

The following provisions are important for AI-enabled arbitration:

Section 7 — Arbitration agreement

The agreement must satisfy the statutory requirements for an arbitration agreement. A digital agreement can raise questions about consent, incorporation of platform terms, and the scope of the dispute-resolution clause.

Section 12 — Grounds for challenge

The provision addresses circumstances that may give rise to justifiable doubts about an arbitrator's independence or impartiality, as well as statutory qualifications and disclosures. These requirements are relevant when assessing human arbitrators and the integrity of AI-supported appointment processes.

Section 18 — Equal treatment of parties

Parties must be treated equally and each must receive a full opportunity to present its case. Undisclosed AI-generated evidence or decisive reasoning that cannot fairly be challenged may raise issues under this provision.

Section 19 — Determination of rules of procedure

The tribunal is generally not bound by the Code of Civil Procedure, 1908 or the Indian Evidence Act, 1872 in the same manner as a court, subject to the Act's requirements. Procedural flexibility does not eliminate the duty to treat parties fairly.

Section 31 — Form and contents of an award

The statutory requirements concerning the form and content of an award remain relevant when AI assists with drafting. A draft produced by software must not be confused with an award validly made by the tribunal.

Sections 34 and 36 — Setting aside and enforcement

Section 34 provides the statutory framework for challenging an award. Section 36 addresses enforcement of qualifying domestic awards, subject to the Act. AI-related defects must be assessed against the relevant legal requirements.

B. Digital and data-protection legislation

The legal framework extends beyond arbitration legislation.

Information Technology Act, 2000: Relevant to electronic records, electronic transactions, and certain cybersecurity-related matters, subject to the Act's scope and exclusions.

Digital Personal Data Protection Act, 2023: Relevant to the processing of digital personal data where its provisions are applicable and in force.

Contract Act, 1872: Relevant to contractual consent, enforceability, and claims involving breach of contractual obligations.

Applicable consumer-protection legislation: May become relevant when arbitration platforms are offered to consumers, depending on the dispute and the governing statutory regime.

A platform's terms of service cannot automatically override mandatory legislation. Its legal obligations must be assessed in light of the applicable statutes, commencement provisions, rules, and contractual arrangements.

C. Can AI independently act as an arbitrator in India?

This question requires careful treatment.

The Arbitration and Conciliation Act, 1996 regulates the constitution and functioning of arbitral tribunals but does not provide a comprehensive, express framework specifically governing autonomous AI arbitrators.

There is therefore an important distinction between:

AI assisting a properly constituted human tribunal.

AI preparing a proposed decision that the tribunal independently reviews.

AI itself issuing a purportedly binding award without meaningful human adjudication.

The first two models can be evaluated within existing procedural and contractual rules. The third raises more fundamental questions concerning the legal identity, appointment, impartiality, responsibility, and authority of the purported decision-maker.

It would be inaccurate to conclude that every AI-assisted award is invalid, or that every autonomous AI award is automatically enforceable. The outcome depends on the statutory framework, the arbitration agreement, and the facts.

8. Civil liability of AI arbitration platform operators

Platform regulation must address not only the validity of awards but also the civil liability of the entities operating the technology.

The remedy depends on the identity of the defendant and the source of the duty. A developer's contractual liability is different from an arbitrator's duties, and an award challenge is different from a damages action against a technology provider.

9. Recommended regulatory framework

An effective regulatory system should preserve the advantages of automation without sacrificing the legal legitimacy of arbitration.

Meaningful human responsibility: Where applicable law or the parties' agreement requires a human arbitrator, AI should remain an assistance tool and the arbitrator should independently assess the material issues and issue the final award.

Risk-based disclosure: Require disclosure of AI use when it materially affects evidence assessment, legal reasoning, or the drafting of an award. Routine administrative automation may require less extensive disclosure.

Independent testing: Assess systems for material bias, accuracy, security weaknesses, and reliability before deploying them in consequential dispute-resolution functions.

Contestability: Allow parties to raise objections to material AI-generated conclusions and provide an appropriate opportunity to respond.

Evidence preservation: Maintain appropriate records of model versions, material inputs and outputs, corrections, and human review so that disputed conclusions can be examined.

Data protection: Use secure systems, access controls, retention limits, and appropriate contractual safeguards for confidential and personal data.

Clear allocation of liability: Define the responsibilities of the platform provider, institutional administrator, arbitrator, and parties without attempting to contract out of mandatory legal obligations.

Proportionate judicial review: Preserve established statutory grounds for challenging awards and refusing enforcement rather than treating the mere use of AI as sufficient to invalidate an award.

These are recommended safeguards, not a claim that every jurisdiction has enacted all of them as mandatory legal requirements.

10. Illustrative hypothetical disputes

The following examples show how existing civil-law principles might apply.

Hypothetical A: AI-generated award contains fabricated authorities

A technology company brings a contractual claim against a supplier. The platform produces a draft award citing three judicial decisions that do not exist. The arbitrator adopts the draft without checking the authorities.

Legal analysis: The fabricated citations may indicate serious deficiencies in the decision-making process. The consequences depend on whether the errors materially affected the award, whether the arbitrator independently assessed the case, and whether a statutory ground for challenge exists.

A court would not necessarily set aside the award solely because a citation was fabricated; the applicable legal test and the impact on the proceedings would matter.

Hypothetical B: Biased AI model favors one party

An online platform uses a proprietary model trained primarily on disputes from one industry. A claimant discovers that the model consistently assigns lower damages to businesses of its size, and the tribunal relied heavily on the model's recommendation.

Legal analysis: The parties may seek appropriate information about the methodology, data limitations, and reliability of the recommendation. If the circumstances establish material procedural unfairness or another recognized ground, the issue may support a challenge.

The mere existence of a statistical disparity is not necessarily sufficient to prove unlawful discrimination or invalidate an award. Further evidence and the applicable legal standard are essential.

Hypothetical C: Confidential documents are disclosed to an external AI provider

During arbitration, a party uploads commercially sensitive contracts and customer records to an AI service that retains the data contrary to the platform's contractual commitments.

Legal analysis: Depending on the facts, this may raise contractual confidentiality, data-protection, security, and damages issues. The affected party may seek containment of the breach and appropriate legal remedies. Regulatory notification obligations may also arise.

This dispute is separate from whether the eventual arbitral award is valid.

11. Advantages and disadvantages of AI arbitration platforms

AI may improve procedural efficiency, particularly in high-volume or document-intensive disputes. However, speed and lower cost do not by themselves establish that a process is fair, legally valid, or suitable for every dispute.

12. Conclusion

Civil law and AI arbitration platform regulation concern the relationship between technological innovation and established principles of arbitration: consent, independence, impartiality, procedural fairness, accountability, and enforceability.

The eight judicial decisions discussed above provide relevant legal foundations, but none should be misrepresented as a direct judicial ruling on the validity of an autonomous AI arbitrator.

The principal regulatory challenge is to determine how existing arbitration principles apply when AI materially influences the adjudicative process. Human oversight, appropriate disclosure, reliable evidence, confidentiality safeguards, and clear allocation of responsibility can help address these risks.

Ultimately, the legal validity of an AI-assisted arbitral award depends not on the mere presence of artificial intelligence, but on whether the agreement, tribunal, procedure, and resulting award comply with the applicable law.

Academic note: This is a comparative legal overview, not jurisdiction-specific legal advice. For a case-law assignment, distinguish direct precedent on arbitration from analogous precedent applied to emerging AI technology.