Ushering in a new era: The Austrian outlook of deployment and regulation of AI in international arbitration

Publications: March 19, 2025

Back to overview

Authors

Related Expert Guides

Introduction

Artificial intelligence (hereinafter AI) in alternative dispute resolution and international arbitration, in particular, is no longer a tale of the distant future but an objective reality that was given legislative shape last summer at the European Union law level.

The adoption of the European Union’s breakthrough and indeed anticipated Artificial Intelligence Act, Regulation (EU) 2024/1689 (hereinafter EU AI Act) has hardly left anyone indifferent. The EU AI Act affects multiple areas of society, and alternative dispute resolution, a substantial part of which is international arbitration, is no exception.[1]

Considering a wide range of diverse approaches and bypassing technical discussions to the defining concept of AI, this article favors the definition of AI expressed in Article 3 (1) of the EU AI Act.

This article will scrutinize the impact of AI and its deployment in international arbitration, as well as the potential problems and controversies arising from using AI in international arbitration through the lens of Austrian regulation.

Deployment and use of AI in international arbitration

Artificial intelligence's significant leap in popularity has led to the deployment of such technologies in international arbitration and an attempt to categorize and understand the limits, benefits, and potential risks of such symbiosis. 
The following classification of the use of AI in international arbitration is noteworthy:

  • case or decision prediction;

  • drafting an arbitration clause or arbitration agreement;

  • choosing arbitrator(s);

  • drafting legal documents and arbitral awards;

  • document review and data analysis.[2]

This list is certainly not exhaustive due to the significant degree of parties autonomy in international arbitration and the uniqueness of AI, which is capable of solving (although not without errors) completely different tasks. For instance, AI may also translate a vast array of documents into different languages, considering the internationality of disputes in arbitration[3]  or AI could be used by arbitration institutions for the case management purposes.[4]  

Another approach to classifying the use of AI in international arbitration is reflected in the Silicon Valley Arbitration and Mediation Center (hereinafter SVAMC) Guidelines, which are divided into three parts depending on the actors using AI:

  • Guidelines for All Participants in Arbitrations;

  • Guidelines for Parties and Party Representatives;

  • Guidelines for Arbitrators.[5] 

These Guidelines are of particular importance because they constitute the first attempt at supplementary international standards level to implement and regulate the use of AI in international arbitration proceedings, taking into account the cornerstone concepts of international arbitration such as: confidentiality, mandatory rules of law, due process, non-delegation of decision-making responsibilities by arbitrators, and etc.

The above-mentioned popularity of using AI in international arbitration is additionally proved by several recent surveys:

  • according to BCLP Annual Arbitration Survey 2023 on the use of Artificial Intelligence in International Arbitration, 90% of respondents were aware that AI tools existed that could perform a range of tasks in international arbitration;[6] 

  • As stated by the Survey of Arbitral Institutions for the ICCA Congress Panel on AI, out of the 11 institutions that responded, 4 indicated they have implemented AI in some form, and all of the responding institutions said they recognize AI’s potential and are considering whether and how to adopt it in the future.[7] 

Consequently, any phenomenon gaining such popularity while presenting obvious risks (they will be addressed later in the article) and being of great benefit to society and its users will inevitably be regulated by the competent authorities.

Regulation of AI in arbitration in EU and Austria

The EU Artificial Intelligence Act

The EU AI Act, significant not only in its scope and size but also in its impact on AI's future fate, has not bypassed the issue of deployment of AI in alternative dispute resolution. The cornerstone of the EU AI Act is the “risk-based approach,” according to which the use of AI systems is divided into categories corresponding to the amount of risk that the AI system can cause to public health, safety, public security, fundamental rights, or the society as a whole.

Pursuant to Recital 61 of EU AI Act, to address the risks of potential biases, errors and opacity, it is appropriate to qualify as high-risk AI systems intended to be used by a judicial authority or on its behalf to assist judicial authorities in researching and interpreting facts and the law and in applying the law to a concrete set of facts. AI systems intended to be used by alternative dispute resolution bodies for those purposes should also be considered to be high-risk when the outcomes of the alternative dispute resolution proceedings produce legal effects for the parties. The use of AI tools can support the decision-making power of judges or judicial independence, but should not replace it: the final decision-making must remain a human-driven activity. The classification of AI systems as high-risk should not, however, extend to AI systems intended for purely ancillary administrative activities that do not affect the actual administration of justice in individual cases, such as anonymisation or pseudonymisation of judicial decisions, documents or data, communication between personnel, administrative tasks.

Wording of Recital 61 such as “alternative dispute resolution bodies” probably refers to arbitrators in international arbitration. A similar position is expressed in the international arbitration community.[8]  

Thus, arbitrators' use of AI in researching and interpreting facts and the law and in applying the law to a concrete set of facts when the outcomes of the alternative dispute resolution proceedings produce legal effects for the parties might be considered as a use of high-risk system under Article 6(2) of the EU AI Act due to the reference to the ADR procedure in Recital 61 and Annex III (8)(a).

Hence, using AI in arbitral proceedings by arbitrators might imply certain obligations under Article 26 of the EU AI Act and even penalties under Article 99 of the EU AI Act for the deployers of high-risk activities in case of non-compliance.

However, it is necessary to designate that the practice of application of the EU AI Act is still far from being formed due to different transition periods outlined in Article 113 of EU AI Act, and these considerations are only theoretical at this stage. 

Austrian national legislation and VIAC developments

The last time the Austrian Arbitration Act, contained in Sections 577 to 618 of the Austrian Code of Civil Procedure (hereinafter ACCP), was updated in 2013, AI systems hardly played the prominent role they do nowadays. Nevertheless, this is not an obstacle to evaluating the existing provisions of the Austrian Arbitration Act for their possible correlation with AI systems.

According to Section 586 (1) ACCP, the parties are free to agree on the number of arbitrators. If the parties have, however, agreed on an even number of arbitrators, then these shall appoint a further person (emphasis added) as chairman. Furthermore, pursuant to Section 588 (1) ACCP, when a person (emphasis added) intends to assume the office of an arbitrator, he (emphasis added) shall disclose any circumstances likely to give rise to doubts as to his impartiality or independence, or which are in conflict with the agreement of the parties.

The aforementioned wording reasonably indicates that the appointment of an AI system as an arbitrator would probably be considered as a violation of the Austrian Arbitration Act since the “personality” criterion would not be met in the event of such an appointment.

In a same vein, the Vienna International Arbitral Centre (hereinafter VIAC) Rules of Arbitration (hereinafter Vienna Rules), implicitly precludes the appointment of artificial intelligence systems as arbitrators. According to Article 6(2) of Vienna Rules, to the extent the terms used in the Vienna Rules refer to natural persons (emphasis added), the form chosen shall apply to all genders. Article 16 (1) of Vienna Rules similarly refers to “persons” regarding parties’ right to nominate arbitrators.

Hence, notwithstanding potential considerations of broad party autonomy in international arbitration, appointing an AI system as an arbitrator would be contrary to the ACCP and inadmissible under the Vienna Rules. These considerations are proved by Recital 61 of the EU AI Act, which states that: “the use of AI tools can support the decision-making power of judges or judicial independence, but should not replace it: the final decision-making must remain a human-driven activity.”  

Meanwhile, it should be noted that electronic technologies as support tool are actively used by the VIAC and are reflected in the Vienna Rules: 

  • according to Article 12 of Vienna Rules statement of claim and exhibits shall be submitted electronicaly and Secretariat shall receive all written communications between the arbitral tribunal and the parties in electronic form;

  • according to Article 30(1) of Vienna Rules having due regard to the views of the parties and the specific circumstances of the case, the arbitral tribunal may decide to hold an oral hearing in person or by other means;

  • according to Article 36(5) of Vienna Rules the Secretariat may send a copy of the award in electronic form. 

Additionally, in 2021 VIAC launched the VIAC portal – an online case management platform aimed at communication and exchange of documents among the VIAC, the parties, and the arbitrators or other third-party neutrals during all types of VIAC proceedings.[9]  More to the point, in 2022, VIAC introduced a new initiative called “Legal Tech Think Tank,” which aimed to collect the expertise and knowledge of the impact of legal tech, cryptocurrency, blockchain, and AI on the alternative dispute resolution landscape.[10]

Furthermore, the Austrian Supreme Court (Oberster Gerichtshof, hereinafter OGH) rendered a decision (Case No.4Ob77/23m) on the question of whether it is appropriate to use AI systems to connect businesses (mostly SMEs that do not have their own legal departments) with matched lawyer their need depending on case issues and practices areas.[11]  This AI system was also able to provide legal research and possible solutions for matched lawyers, but lawyers were not bound by this advice of AI. In turn, the lawyer found by the AI, after performing the service, billed the client, 25 percent of which was withheld by the system. The Austrian Bar filed a lawsuit to prohibit the use of this system on several grounds, such as confidentiality, unfair competition, and the general prohibition of non-lawyers providing legal services. The Austrian Supreme Court disagreed with the Austrian Bar's position and did not prohibit the use of an AI system with the described functionality. However, it declared that the AI system's imposition of such a fee is inadmissible.

Even though the case was not directly related to international arbitration, the Austrian Supreme Court's decision demonstrates how extensively artificial intelligence has been incorporated into the legal industry. At the same time, such a model could be realized when parties search and select arbitrators for their disputes.
These developments, initiatives and case law indicate that the trend towards using artificial intelligence in arbitration as an auxiliary tool is not unnoticed in Austria and is being actively implemented and researched, including at the VIAC level.

Risks and concerns 

Analyzing the experience of using AI in international arbitration, it is necessary to mention the risks associated with its use.

First risk is so-called “hallucination” of AI systems.[12] Because artificial intelligence systems are configured to answer, if not all, then many user inquiries in quite a limited amount of time, there are sometimes cases when there is a lack of information (for example, if it is confidential), when the artificial intelligence generates a smooth-sounding answer that is entirely or partially a fiction. This problem is highly feasible, as already, in some cases, the national courts have dealt with situations where parties have cited non-existent case law generated by AI.[13] This case illustrates the need for lawyers or arbitrators to double-check artificial intelligence systems' results, as they are still far from perfect.

Second risk is artificial intelligence's involuntary bias, which may be reproduced in connection with an attempt by artificial intelligence to analyze information about existing trends in arbitration or law. This problem is particularly relevant to international arbitration since confidentiality is a cornerstone of international arbitration, and the information that artificial intelligence needs to reach the correct conclusion may simply not be available due to its confidentiality. Additionally, problem of bias of AI addressed by SVAMC Guidelines in connection with search and appointment of individuals as arbitrators, experts, counsel, or any other roles in connection with arbitrations.[14]

Third problem concerns any types of evidence in international arbitration proceedings, since AI systems became advanced in creation forgeries of documents, audio, photo and video content. The differences between original evidence and forgeries created by artificial intelligence are sometimes difficult to distinguish. In this regard, should be supported the initiative set out in the SVAMC Guidelines on the right of arbitral tribunal to sanctioned party by any measures available under the governing law and the applicable arbitration rules or the lex arbitri (such as, for example, striking the evidence from the record, or deeming it inadmissible).[15]

The fourth risk of confidentiality breach will always accompany the use of artificial intelligence systems, including in international arbitration, where the privacy issue is particularly acute. Presuming that any AI system can glitch and leak private information at any time, arbitrators, party representatives, and arbitral institutions should be particularly cautious about transmitting confidential information for processing by AI.

The importance of confidentiality in international arbitration from the perspective of the use of AI tools is also raised in Principles Supporting the Use of AI in Alternative Dispute Resolution Developed by The American Arbitration Association-International Centre for Dispute Resolution (hereinafter AAA-ICDR): “Safeguarding sensitive data is critical in ADR, as in business and the law. The integration of AI should not compromise this principle. Preventing unauthorized access, leakage, or misuse of confidential data is essential. Special care is needed with large datasets, opaque machine learning models, and uncertain data protocols”.[16]

Evidently, when two large-scale phenomena, such as artificial intelligence and international arbitration, collide, the list of risks and challenges of their symbiosis mentioned above is not exhaustive and may change over time, but the problems outlined are already crucial for the international arbitration community.

Conclusion

In a nutshell, the use of AI in international arbitration will grow exponentially. The deployment of AI systems by arbitrators, party representatives, and arbitral institutions poses many regulatory and ethical questions to which the international arbitration community, international organizations, and competent national authorities have yet to find answers. At this stage, artificial intelligence systems in international arbitration are perceived only as high-speed and often efficient, but still auxiliary tools that cannot replace or supplant arbitrators, party representatives, or arbitral institutions.

Taking Austria as an example, we can conclude that the VIAC pays due attention to the issue of artificial intelligence. Even more, given that the Austrian Supreme Court (OGH) has already spoken about artificial intelligence (although not in the context of arbitration), the issue of using artificial intelligence in arbitration will be further developed not only at EU law or Austrian national law level, but also in Austrian case law.
 

Ressources

  1. Regulation (EU) 2024/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act) – Text with EEA relevance. Please see here: https://eur-lex.europa.eu/eli/reg/2024/1689/oj .
  2. For a more detailed assessment of the possible applications of AI in arbitration see, Shih, Sean and Chang, Chin-Ru, The Application of AI in Arbitration: How Far Away are we from AI Arbitrators? (May 31, 2024). Contemporary Asia Arbitration Journal, Vol. 17, No. 1, pp. 69-90, Available at SSRN: https://ssrn.com/abstract=4849614
  3. See Para. 5.2. of ICC Arbitration and ADR Commission Report on Leveraging Technology for Fair, Effective and Efficient International Arbitration Proceedings Available at: https://iccwbo.org/news-publications/arbitration-adr-rules-and-tools/icc-arbitration-and-adr-commission-report-on-leveraging-technology-for-fair-effective-and-efficient-international-arbitration-proceedings/
  4. For a more detailed assessment of the possible use of AI in case management in arbitration see, Ahmet Cemil Yıldırım, The use of technology in case management in international investment arbitration: a realistic approach, Arbitration International, Volume 40, Issue 2, June 2024, pp. 233–250, Available at: https://doi.org/10.1093/arbint/aiae010.
  5. See SVAMC Guidelines on the Use of Artificial Intelligence in Arbitration 1st edition 2024 Available at: https://svamc.org/svamc-publishes-guidelines-on-the-use-of-artificial-intelligence-in-arbitration/
  6. See BCLP Arbitration Survey 2023 on the use of Artificial Intelligence in International Arbitration Available at: https://www.bclplaw.com/en-US/events-insights-news/bclp-arbitration-survey-2023.html
  7. See Maxim Osadchiy & Erika Santini, Are Arbitral Institutions Using Artificial Intelligence? The State of Play in Adopting AI, KLUWER ARB. BLOG (May 8, 2024), Available at: https://arbitrationblog.kluwerarbitration.com/2024/05/08/are-arbitral-institutions-using-artificial-intelligence-the-state-of-play-in-adopting-ai/
  8. See Maxi Scherer, We Need to Talk About … the EU AI Act!, KLUWER ARB. BLOG (May 27, 2024), Available at: https://arbitrationblog.kluwerarbitration.com/2024/05/27/we-need-to-talk-about-the-eu-ai-act/
  9. For more, see: https://viac.eu/en/arbitration/viac-portal
  10. For more, see: https://viac.eu/en/news/viac-getting-tech-savvy-viac-launches-legal-tech-think-tank
  11. Austrian Supreme Court 4Ob77/23m of 27 June 2023, Available at: https://www.ris.bka.gv.at/Dokument.wxe?Abfrage=Justiz&Dokumentnummer=JJT_20230627_OGH0002_0040OB00077_23M0000_000&Suchworte=RS0079640
  12. See, SVAMC Guidelines on the Use of Artificial Intelligence in Arbitration p.16.
  13. Mata v. Avianca, Inc., 678 F. Supp. 3d 443 (S.D.N.Y. 2023) Available at: https://casetext.com/case/mata-v-avianca-inc-3
  14. See, SVAMC Guidelines on the Use of Artificial Intelligence in Arbitration p.16.
  15. Ibid. p.19.
  16. Principles Supporting the Use of AI in Alternative Dispute Resolution, American Arbitration Association (Nov. 2023) Available at: https://go.adr.org/rs/294-SFS-516/images/Principles%20Supporting%20the%20Use%20of%20AI%20in%20Alternative%20Dispute%20Resolution.pdf
Back to overview