Page 53
CENTRAL ASIAN JOURNAL OF ACADEMIC
RESEARCH
Volume 3, Issue 6, Part 4 Iyun 2025
www.in-academy.uz
BALANCING EFFICIENCY AND FAIRNESS: AI IN
INTERNATIONAL DISPUTE RESOLUTION
Sabina Almatova
Master’s student at Tashkent State Law University, specializing in
"International Arbitration and Dispute Resolution"
https://doi.org/10.5281/zenodo.15798942
ARTICLE INFO
ABSTRACT
Qabul qilindi: 25-Iyun 2025 yil
Ma’qullandi: 28-Iyun 2025 yil
Nashr qilindi: 30-Iyun 2025 yil
This article explores the multifaceted role of artificial
intelligence (AI) in the context of international
commercial arbitration. It discusses how AI is being
integrated at various procedural by both parties and
arbitral institutions. The author examines the benefits of
AI, such as cost reduction, increased efficiency, and
enhanced legal analysis, as well as its challenges,
including concerns over transparency ("black box"
issues), bias, and legal legitimacy of AI-assisted decisions.
Emphasis is placed on the need for ethical frameworks,
transparency standards, and human oversight to ensure
that technological advancement does not compromise
due process or parties' trust in arbitration.
KEYWORDS
Artificial
Intelligence,
International
Arbitration,
Dispute
Resolution,
Legal
Technology, Black Box Issues,
Legal Ethics, Digital Justice.
Nowadays, AI is increasingly being introduced into various fields of activity. The legal
industry, including arbitration, is no exception to this trend. AI is computer systems capable of
performing tasks that require "human" thinking (recognizing speech, processing natural
language, learning from data)
1
. In other words, AI represents the ability of computer systems
to perform tasks that require human intelligence: information analysis, learning from
experience, decision making. Modern AI systems, especially those based on machine learning
and neural networks, are capable of processing huge amounts of data, identifying complex
relationships and making predictions with high speed and accuracy. In the legal field, AI is
already used for such tasks as: automated translation of documents, intelligent search in
databases of legislation and decisions, analysis of contracts for risks and typical errors,
forecasting dispute outcomes based on precedents and even generating draft legal acts based
on specified parameters. Today, in the arbitration process, AI can act as an auxiliary tool at
different stages. Parties to commercial disputes are using AI tools at various stages of the
proceedings, and arbitrators and lawyers are using algorithms for document analysis,
enforcement, and case management. This is part of the overall digital transformation, with the
1
Odongo, R. (2021). AI (AI) in International Arbitration. Nairobi Center for International Arbitration.
Page 54
CENTRAL ASIAN JOURNAL OF ACADEMIC
RESEARCH
Volume 3, Issue 6, Part 4 Iyun 2025
www.in-academy.uz
introduction of electronic claim portals, ODR systems, and smart contracts
2
. For example, in
India, the Supreme Court has formally allowed arbitration agreements to be concluded by
email and amended the Arbitration Act to allow “electronic means” for the formation of such
agreements. These changes highlight that AI and digital technologies are already penetrating
the commercial arbitration sphere, promising to improve the efficiency of procedures and
expand access to jurisdiction. This trend certainly has a number of advantages. Firstly, AI
significantly speeds up routine processes and reduce costs. Automation of repetitive
operations such as document sorting and analysis, case management, and translation of texts
makes it possible to process large volumes of information without tedious manual work.
Research has shown that the use of AI for legal research and document review can reduce the
time required to complete tasks from months and years to seconds
3
. Machine learning already
enables specialized programs to process and structure large amounts of information in
seconds, select relevant materials, and even generate drafts of procedural documents. This
significantly reduces the workload of arbitrators and representatives of the parties, allowing
them to focus on developing a legal position and strategic aspects of the case. In essence, such
AI tools perform the tasks of a “junior lawyer,” while acting with greater speed and accuracy.
As a result, dispute resolution is significantly accelerated: algorithms are able to generate
draft decisions faster than a person can do manually. This reduces the accumulation of cases,
reduces arbitration costs for the parties, and makes the procedure more affordable
4
. AI has
already been integrated into leading commercial legal platforms such as Westlaw and
LexisNexis, and its use is increasingly recommended for widespread use in arbitration
practice in order to improve efficiency and reduce costs. In addition, AI contributes to a
deeper and more accurate legal analysis. Modern algorithms are able to process complex,
multi-layered data, identifying relationships and patterns that might otherwise go unnoticed
by humans. For example, intelligent systems effectively identify key arguments contained in
the parties’ documents and automatically find relevant or previously missed legal precedents.
Using machine learning, programs learn from thousands of previous decisions and can predict
the outcome of disputes, which helps parties build the right strategy, since predicting the
outcome allows them to adjust their position and arguments at an early stage. Finding
relevant legal norms, doctrine, and precedents is a labor-intensive part of preparing for a case.
Platforms with AI tools can speed up this process: for example, Jus Mundi has launched Jus -
AI, an assistant that answers questions based on a vast database of arbitration decisions and
documents. Lexis+AI and Westlaw Edge, legal information products, integrate GPT models to
understand a lawyer’s natural language query and return accurate results with sources. They
also provide analytics: for example, Westlaw Edge can analyze the practice of a specific judge
or arbitrator on request, showing the inclinations in decisions. For arbitration, where it is
often necessary to study foreign laws and precedents, such tools are especially useful. Of
2
Comprés , T. (2025, May 6). Global Miami magazine. Global Miami Magazine - a Hub for Miami's Booming
International Business Community.
https :// globalmiamimagazine . com /2025/05/06/ the - role - of - technology - in -
3
Odongo, R. (2021). AI (AI) in International Arbitration. Nairobi Center for International Arbitration.
4
Bhumika Indulia . (2025, January 8). Arbitration in the Era of AI: What the Future Holds. SCC Times.
https://www.scconline.com/blog/post/2025/01/08/arbitration-in-the-era-of-ai-experts-corner/
Page 55
CENTRAL ASIAN JOURNAL OF ACADEMIC
RESEARCH
Volume 3, Issue 6, Part 4 Iyun 2025
www.in-academy.uz
course, AI cannot yet completely replace a legal researcher – it is not trusted without
verification – but in routine queries it saves a lot of time.
Reducing time costs with the help of AI can also be seen in the process of selecting an
arbitrator. By analyzing a large volume of biographical data and rulings, AI tools can help
select candidates with the necessary qualifications, while taking into account their possible
biases. Here, it is not exactly AI that comes to the fore, but rather databases and analytics. For
example, the Arbitrator Intelligence project collects feedback and information on the behavior
of arbitrators in different cases to help parties make informed decisions
5
. The platform does
not provide a ready-made answer, but systematizes the information: how many cases the
arbitrator has handled, the tendency to grant motions for evidence, the average duration of
trials under his or her chairmanship, etc. In the future, AI may be able to recommend
arbitrators for a specific dispute profile, taking into account their experience, and also predict
how a particular candidate may treat certain arguments. However, it is necessary to take into
account that such systems are created in compliance with confidentiality: information about
arbitrators should not contain references to specific cases. This is an example of how AI tools
can increase the transparency of the arbitration services market by reducing dependence on a
limited number of renowned arbitrators and giving a chance to new professionals. However, it
is important to note that excessive automation of selection can lead to a decrease in individual
assessment and unintentional discrimination. In this regard, AI in this case should be used as
an auxiliary tool, and the final decision should remain with the parties or the relevant
appointing authority. Nevertheless, even in this form, modern technologies significantly
increase transparency and accountability in the procedure for the formation of the arbitral
tribunal, strengthening trust in arbitration as a dispute resolution mechanism.
Another advantage of using AI in international arbitration, which is important for the
parties to the dispute, may be predicting the outcome of the case. Thus, the most complex and
intellectually rich form of using AI in arbitration is predicting the outcome of a dispute or, at
least, assessing the probability of success of a particular position. For example, ArbiLex is
aimed at arbitration: it analyzes uploaded legal documents: claims, responses to claims,
contracts, decisions on similar cases – and, based on embedded machine learning algorithms,
produces a probabilistic forecast of how the case may be resolved
6
. This is especially valuable
for the parties at the preparation stage: based on the forecast, they can decide whether to
continue the dispute, whether to conclude a settlement agreement and on what terms.
Another platform, Case Crunch , was initially used to predict banking disputes in the UK, but
the technology can be expanded. Of course, no such system guarantees 100% accuracy – the
arbitral tribunal may decide otherwise – but the high success rate demonstrates that AI
systems can detect patterns that are invisible to humans. The risk, however, is that such
predictions may introduce some bias: for example, if an arbitrator learns that “the machine
thinks the claimant is 90% likely to win,” this may subconsciously influence their decision.
Therefore, such tools are usually used only by parties for internal risk assessment or by
5
Arbitrator Intelligence | About. Arbitratorintelligence.vercel.app .
https :// arbitatorintelligence . vercel . app / about
6
Abdurakhmanova, N. (2024). LEGAL REGULATION OF THE USE OF AI IN ONLINE ARBITRATION:
INTERNATIONAL AND NATIONAL ASPECTS. Izhtimoiy-humanitarian fan larning dolzarb muammolari / Actual
problems of social and humanitarian sciences / Actual Problems of Humanities and Social Sciences., 4(8).
Page 56
CENTRAL ASIAN JOURNAL OF ACADEMIC
RESEARCH
Volume 3, Issue 6, Part 4 Iyun 2025
www.in-academy.uz
arbitral institutions for statistical analysis, but not by the arbitrators themselves when making
an award.
Noting the use of modern technologies in international arbitration, it is worth
emphasizing that AI can be used not only by the parties, but also by arbitrators. AI-based tools
facilitate the process of formulating a decision, helping to build its logical structure.
Generative models can turn an arbitrator’s sketches and notes into a coherent text, which is
then, if necessary, corrected by a person. The DLA Piper study notes that generative AI is
already being used to create “essays” – for example, converting transcripts or notes into draft
statements - and editing such drafts is much faster than writing from scratch
7
. The arbitrators
themselves note that AI-based tools simplify the process of writing a decision and eliminate
the need for countless checks, allowing them to concentrate on the substantive part. In other
words, the system has become a kind of multifunctional proofreader that automatically
checks the integrity of the text, style and correctness of references, while a person formulates
the final conclusions and justifications. As a result, the structure of the arbitration act
becomes more logical, since even complex arguments are initially formulated in a clear
sequence, and the arbitrator only refines the nuances. This approach reduces the risk of
missing key points and ensures the integrity of the arbitration decision, which is certainly an
advantage of using modern technologies.
Despite a number of advantages of using AI in international arbitration, this
phenomenon still carries certain risks, considering which, first of all, it is necessary to
mention the issue of the legitimacy of the arbitration award when using AI. Thus, the
widespread use of AI may lead to new arguments when challenging arbitration awards. For
example, a party may apply to the court with a statement that the dispute was actually
resolved not by a person, but by a program, which means that the procedure for forming the
arbitration panel prescribed by law was not followed. Such a situation may be a consequence
of the lack of empathy and human understanding in AI. Complex commercial conflicts often
involve emotional and human aspects. Arbitrators often use a comprehensive assessment
based on experience, a sense of justice, and sometimes even empathy for the bona fide party.
AI, on the other hand, lacks a real understanding of human values. Research notes the
importance of emotional intelligence and direct human interaction in dispute resolution,
which is inaccessible to AI
8
. In other words, an AI-based tool may make a “cold” decision,
which reflects another potential problem – the risk of machine bias. This drawback is
associated not only with the development of the technology in a particular state, but also
directly with the machine learning process. For example, if in the past arbitrators
systematically made decisions against a certain category of companies, the AI may consider
that belonging to this category is a factor for losing, and “advise” the corresponding decision.
In addition, speaking about a certain “bias”, it is also worth noting the risks of politically
biased decisions made by AI. Thus, due to interstate relations, AI may make a decision against
an “unfriendly” jurisdiction. Such situations, in turn, may become grounds for the cancellation
of the arbitration award. Although there are no such precedents to date, it is worth noting that
7
IA
Meets
AI
–
Rise
of
the
Machines
|
DLA
Piper.
(2023).
DLA
Piper.
8
Arbitration Tech Toolbox: AI as an Arbitrator: Overcoming the “Black Box” Challenge? - Kluwer Arbitration Blog.
(2024, August 23). Kluwer Arbitration Blog.
https://arbitrationblog.kluwerarbitration.com/2024/08/23/arbitration-tech-
toolbox-ai-as-an-arbitrator-overcoming-the-black-box-challenge/
Page 57
CENTRAL ASIAN JOURNAL OF ACADEMIC
RESEARCH
Volume 3, Issue 6, Part 4 Iyun 2025
www.in-academy.uz
if the arbitrator made a decision following the conclusion of an analytical program based on
AI, the dissatisfied party will try to use this situation to cancel the decision.
One of the key shortcomings of AI today is the “black box” problem. “Black boxes” are AI
systems whose internal mechanisms remain a mystery to their users
9
. In other words, there is
no transparency in the operation of the algorithms: the user can only see the input data and
the result of their consideration, while the process of arriving at a certain conclusion, i.e. what
factors were taken into account by the algorithm when making a decision, remains
undisclosed. Many modern AI systems, especially deep neural networks, operate according to
complex non-linear models that are difficult to explain in understandable terms. In the
context of arbitration, this can be seen as a serious obstacle: the arbitration decision must be
motivated, therefore, the arbitration panel must understand the reasons for the interim
decisions they make. If it is unclear why the AI recommended rejecting a particular piece of
evidence, such advice cannot be relied upon. Moreover, the lack of transparency undermines
trust: even if the algorithm is statistically correct 95% of the time, the party that loses because
of the “AI advice” will always doubt whether there is an error in the remaining 5%
10
. In this
regard, arbitrators avoid using “black” algorithms where justification is required.
Confidentiality issues are of no small importance in arbitration. In this context, another
problem may arise: for example, if a party uploaded case materials to an external AI service
for translation or analysis, the other party may be outraged that its confidential data was
actually transferred to a third party - a software developer, which, in turn, is a violation of the
confidentiality agreement. In addition, ethical aspects such as liability for the leakage of
confidential data have not yet been studied. Thus, the question still remains open: who is
responsible for the dissemination of data, for an incorrect decision due to the bias of the AI
system: the arbitration panel or the developers of the AI software? These legal aspects are still
outside traditional regulation and create an area of uncertainty.
The use of AI in international arbitration promises significant benefits: reduced
transaction costs, improved quality of legal analysis, automation of routine tasks, acceleration
of procedures, etc. At the same time, as rightly noted in the Chartered Institute of Arbitrators
Guide to the Use of AI in Arbitration (2025), with each new level of automation new ethical
challenges appear that cannot be ignored without compromising the legitimacy of the
arbitration institution itself
11
. AI is a system capable of performing intellectual tasks that
require human participation: text analysis, information processing, prediction of
consequences, etc. In legal practice, AI means are used to search for and analyze precedents,
draft documents, assess risks, predict case outcomes, and, more recently, in arbitration
proceedings. Systems such as ROSS Intelligence, LexisNexis, Kira Systems are currently
actively used in law firms and arbitration institutions to automate various stages of legal
analysis. However, despite a number of advantages, the current level of development of AI,
especially language models, is far from perfect. Generative models are notorious for their
tendency to invent convincing-sounding but false information if they have gaps in their
9
Kosinski, M. (2024, October 29). What is black box AI (AI)? IBM.
https://www.ibm.com/think/topics/black-box-ai
10
Arbitration Tech Toolbox: AI as an Arbitrator: Overcoming the “Black Box” Challenge? - Kluwer Arbitration Blog.
(2024, August 23). Kluwer Arbitration Blog.
https://arbitrationblog.kluwerarbitration.com/2024/08/23/arbitration-tech-
toolbox-ai-as-an-arbitrator-overcoming-the-black-box-challenge/
11
Chartered Institute of Arbitrators ( CIArb ),
Guideline on the Use of AI in Arbitration
,
Page 58
CENTRAL ASIAN JOURNAL OF ACADEMIC
RESEARCH
Volume 3, Issue 6, Part 4 Iyun 2025
www.in-academy.uz
knowledge. This makes them potentially unreliable sources for decision-making, especially in
cases that require strict adherence to the rule of law, accurate interpretation of legal concepts,
and consideration of the unique context of the dispute. Moreover, even if AI is able to formally
formulate a reasoned decision, the problem of the lack of genuine legal judgment and
professional responsibility remains. Arbitration, unlike an automated procedure, presupposes
not only the technical application of the rule, but also consideration of fairness, good faith,
balance of interests of the parties, and other categories that require human assessment. Thus,
the successful use of AI requires the development of international standards and an ethical
framework governing its use, as well as ensuring a sufficient level of human oversight and
participation. Developments in technology open up new horizons for arbitration, but these
opportunities cannot be considered without appropriate regulatory and institutional support.
Therefore, the task of the international legal community is not simply to introduce AI into
arbitration, but to do so thoughtfully, taking into account all legal, ethical and procedural
consequences
References:
1.
Odongo, R. (2021). AI (AI) in International Arbitration. Nairobi Center for International
Arbitration. https://ncia.or.ke/wp-content/uploads/2021/08/ARTIFICIAL-INTELLIGENCE-
AI-IN-INTERNATIONAL-ARBITRATION.pdf
2.
Comprés , T. (2025, May 6). Global Miami magazine. Global Miami Magazine - a Hub for
Miami's Booming International Business Community. https :// globalmiamimagazine . com
/2025/05/06/ the - role - of - technology - in - international - arbitration /
3.
Bhumika Indulia . (2025, January 8). Arbitration in the Era of AI: What the Future Holds.
SCC Times. https://www.scconline.com/blog/post/2025/01/08/arbitration-in-the-era-of-ai-
experts-corner/
4.
Arbitrator Intelligence | About. Arbitratorintelligence.vercel.app . https ://
arbitatorintelligence . vercel . app / about
5.
Abdurakhmanova, N. (2024). LEGAL REGULATION OF THE USE OF AI IN ONLINE
ARBITRATION: INTERNATIONAL AND NATIONAL ASPECTS. Izhtimoiy-humanitarian fan
larning dolzarb muammolari / Actual problems of social and humanitarian sciences / Actual
Problems
of
Humanities
and
Social
Sciences.,
4(8).
https://doi.org/10.47390/spr1342v4i8y2024n47
6.
IA Meets AI – Rise of the Machines | DLA Piper. (2023). DLA Piper.
https://www.dlapiper.com/en/insights/publications/arbitration-matters/2023/ia-meets-ai-
rise-of-the-machines
7.
Arbitration Tech Toolbox: AI as an Arbitrator: Overcoming the “Black Box” Challenge? -
Kluwer
Arbitration
Blog.
(2024,
August
23).
Kluwer
Arbitration
Blog.
https://arbitrationblog.kluwerarbitration.com/2024/08/23/arbitration-tech-toolbox-ai-as-
an-arbitrator-overcoming-the-black-box-challenge/
8.
Kosinski, M. (2024, October 29). What is black box AI (AI)? IBM.
https://www.ibm.com/think/topics/black-box-ai
9.
Arbitration Tech Toolbox: AI as an Arbitrator: Overcoming the “Black Box” Challenge? -
Kluwer
Arbitration
Blog.
(2024,
August
23).
Kluwer
Arbitration
Blog.
https://arbitrationblog.kluwerarbitration.com/2024/08/23/arbitration-tech-toolbox-ai-as-
an-arbitrator-overcoming-the-black-box-challenge/
Page 59
CENTRAL ASIAN JOURNAL OF ACADEMIC
RESEARCH
Volume 3, Issue 6, Part 4 Iyun 2025
www.in-academy.uz
10.
Chartered Institute of Arbitrators (CIArb), Guideline on the Use of AI in Arbitration,
https://www.ciarb.org/media/m5dl3pha/ciarb-guideline-on-the-use-of-ai-in-arbitration-
2025-_final_march-2025.pdf
