1  Online dispute resolution (‘ODR’) is a well-recognized concept nowadays, but its definition is still debated. In broad terms, ODR includes e-Courts and the many forms of alternative dispute resolution (‘ADR’) including negotiation, mediation, and arbitration (Wang 2009, 23). It is argued that ADR is more commonly used than courts as ‘appropriate dispute resolution’ (Menkel-Meadow, 2018). ODR is an extension of ADR to appropriate dispute resolution that uses electronic communications such as emails and other electronic means, providing an easier and cheaper dispute resolution service compared with traditional methods. Currently, ODR services still require a great degree of human interaction such as a human arbitrator or mediator. Whilst ODR can be used in most civil and commercial disputes, those involving Internet-related transaction cases are most suitable due to the accessibility of electronic evidence (Wang, 2017a, 8). With the additional use of artificial intelligence (‘AI’), ODR services may be further assisted or may even become entirely automatic. Although there is no uniform definition of AI, the basic concept is to ‘make appropriate generalizations in a timely fashion based on limited data’ (Kaplan, 2016, 5). The European Commission Proposal for a Regulation laying down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act) and amending certain Union Legislative Acts (2021) (‘Proposal for AI Regulation’) states that an artificial intelligence system (‘AI system’) means:

2  This entry discusses the development of ODR services. It shares insights into how to use them to resolve disputes successfully, in particular online consumer disputes. It considers self-regulatory and regulatory developments concerning ODR and also analyses the merits and limitations of the European Union (‘EU’) ODR Portal in terms of functionality. It seeks to aid the understanding of the legal and technological challenges of ODR services in the age of artificial intelligence and promote the collaboration of private and public justice to improve legal certainty and access to justice.

3  Private ODR services—ie private independent companies offering ODR services—emerged in the late 1990s, and public ODR services—ie not-for-profit public organizations offering ODR services—appeared in the late 2000s. These organizations target regional consumer services and court filing services. For example, SquareTrade used to provide ODR services to eBay but it ceased its ODR service and instead has provided other products such as warranties in recent years (see SquareTrade Warranties on eBay, 2021). In Germany, Online Schlichter provided online mediation services between 2009 and 2019 and was considered as one of most successful German ODR platforms. It received 859 cases in 2012, 1142 cases in 2013, and around 1500 cases in 2014 (see Online Dispute Resolution for Low Value Civil Claims, 2015, 15). Other private ODR services have continued to flourish. For example, Cybersettle provided its service to its business associations with the American Arbitration Association (‘AAA’) (Miller-Moore and Jennings 2007, 5). Based in London, the Centre for Effective Dispute Resolution (‘CEDR’) provides online mediation services using digital platforms (see CEDR website). In 2014, a private entity ‘Youstice’, established in Slovakia, launched an ODR service dealing with customer complaints about online shopping retailers; it uses ideas such as web-based applications for framing arguments broadly similar to other private ODR entities on the market (see Youstice website).

4  Sometimes not-for-profit ODR services may also be semi-public ODR services. For example, in 2007 the Netherlands Ministry of Justice and Security for the Dutch Legal Aid Board and the Hague Institute for Innovation of Law (‘HiiL’)—established in 2005, formerly called the Hague Institute for the Internationalisation of Law—partnered with ODR service providers such as Modria to provide the ODR platform known as Rechtwijzer 2.0 (see HiiL, ‘The Technology of Access to Justice: Rechtwijzer 2.0’, 2021). This ODR system is designed to help parties resolve disputes by diagnosing problems, facilitating resolution via asynchronous dialogue, and question and answer-based framing (‘Q&A’). At the negotiation stage, this ODR system provides automated legal guidance, based on the answers given in a Q&A session. If negotiation fails, it moves to online mediation or arbitration. In the United Kingdom (‘UK’), there are also well-established ADR or ODR services in specific industries such as the Financial Ombudsman Service (since 2000) and Nominet domain names dispute service (since 1996).

5  Since 2011 some new ODR services with distinctive features and special market penetration have also emerged—in particular public ODR services. In the UK, the Traffic Penalty Tribunal provides an online platform for drivers to appeal against penalty charge notices issued by local authorities in England and Wales (see Traffic Penalty Tribunal). In the EU, the single-entry ODR platform has been accessible to consumers and traders since 15 February 2016 (see European Commission, Resolve your consumer complaint). On 1 June 2017 the Civil Resolution Tribunal in Canada began resolving small claims disputes with a value of less than $5,000 (see Civil Resolution Tribunal). There are several stages in which the online tribunal operates: negotiation—to help users explore possible solutions and provide them with the tribunal’s online negotiation platforms; mediation—to assist the parties with a tribunal case manager; and adjudication—to make a final and binding decision by an adjudicator via an online platform and through video conferencing where necessary (Online Dispute Resolution for Low Value Civil Claims, 13).

6  Both private and public ODR services are crucial to improving the public’s access to justice because they can provide appropriate specialist services to different groups of users. For example, the EU single-entry ODR platform provides specialist services (for disputes arising from cross-border online purchases or services) which are for a specific group of users—consumers and traders in the EU. Internal ODR services may be embedded in online trading platforms such as eBay’s Dispute Resolution Centre. Other independent private ODR services in the EU may fill the gaps in public ODR services, targeting domestic consumer disputes or other types of Internet-related disputes, or even offline disputes.

7  It is worth noting that courts often encourage mediation at the first stage to avoid any judicial burden and improve the efficiency of dispute resolution. Courts may penalize litigants who unreasonably refuse mediation (Halsey v Milton Keynes General NHS Trust, 2004). When using a private ODR service, it is also common practice to use negotiation and mediation first before entering into arbitration. For example, in 2006, the AAA and Cybersettle announced a strategic alliance that would provide both companies’ clients with the opportunity to use the dispute resolution services of both companies exclusively, offering clients resolution by mediation first, and if that failed, by arbitration (Miller-Moore and Jennings 2007, 8). The professional regulations of the AAA can be integrated into the self-regulation of private ODR services to promote a harmonized standard of ODR practice, whilst Cybersettle can contribute its expertise in private practices (Wang, 2017a, 28).

8  In the last decade professional organizations have made progress in self-regulatory ODR developments. In 2009, the China International Economic and Trade Arbitration Commission (‘CIETAC’) launched the world’s first Online Arbitration Rules, which were subsequently revised in 2014 (CIETAC Online Arbitration Rules, 2014). More recently other arbitration organizations have focused their developments on online arbitration and mediation. For example, the AAA established the division ‘AAA Mediation.org’, offering clients online dispute resolution (see AAA Mediation.org Mediation Procedures, 2013). The American Arbitration Association International Centre for Dispute Resolution (‘AAA-ICDR’) (International Centre for Dispute Resolution: American Arbitration Association (AAA)) Zoom Virtual Hearing Managed Services also provide an online hearing service in mediation and arbitration to AAA clients, and to members of the public who wish to use a neutral organization to manage their virtual hearing. This service is designed to increase user confidence of a fair process (see AAA Virtual Hearing Managed Services). In 2020, the AAA-ICDR brought out its Order and Procedures for a Virtual Hearing via Videoconference; Virtual Hearings Guide for Arbitrators and Parties Utilising Zoom; and Videoconference Mediation Guide for Parties and Mediators (see AAA/ICDR ‘Virtual Hearings’). Other rules on arbitration and mediation rules have been updated to take into consideration electronic means and usage. For example, the London Court of International Arbitration (‘LCIA’) updated its arbitration and mediation rules in 2020 (Updates to the LCIA Arbitration Rules and the LCIA Mediation Rules, 2020). The Australian Centre for International Commercial Arbitration (‘ACICA’) also issued an Online Arbitration Guidance Note in 2020 and a Draft Procedural Order for the Use of Online Dispute Resolution Technologies in 2016 (see ACICA Practice & Procedures Toolkit).

9  Arbitration, another possible form of private justice, is popular in resolving international commercial disputes. Online arbitration is the most formal type of ODR. Choosing arbitration, whether conducted in person or online, usually means giving up the option of going to court. Scholars have suggested that private/arbitral and public/judicial partnerships relieve the burden on the judiciary and improve access to justice (see Haydock and Henderson, 2002, 141–98). For example, in June 2018 China established a one-stop dispute resolution centre with diversified settlement mechanisms for international commercial disputes including arbitration, and mediation using the China International Commercial Court (‘CICC’) (see CICC website). The CICC will make it more convenient for parties involved in mediation and arbitration, because they can apply to the CICC for the revocation or implementation of an international commercial arbitral award. However, judgments from the CICC, which is part of the Supreme People’s Court (‘SPC’), are final, without the possibility of appeal, and currently with limitations on transnational enforcement (Chaisse and Xu, 2021, 20). Transnational enforcement on judicial judgments is a universal challenge. Even if China ratifies the Hague Convention on Choice of Court Agreements in the future, the enforcement issue in a foreign country may not be completely resolved.

10  Although in theory ODR can be used to resolve disputes both in and out of court, one of the most common reasons for using ODR services is online consumer disputes. If consumers shop on an established online shopping platform such as eBay or Amazon, the embedded dispute resolution service on that platform, such as eBay Resolution Centre, can be used when disputes arise. eBay was established in 1995 and has become one of the world’s largest online marketplaces providing trading platforms. SquareTrade was set up in 1999, providing online dispute resolution services. Like eBay, SquareTrade was an independent private company which shared eBay’s aim of promoting customer confidence in doing business and using services online. It is well-known that the eBay e-trust—ie trust in digital environments—strategies are designed to make customers comfortable when buying and selling online so as to maximize the number of sellers and buyers attracted to its online marketplace (see Taddeo, 2009, 23–35). The trust building measures of eBay include: (1) the mutual rating system of trade satisfaction; (2) identity verification; (3) secure online payment services like PayPal or Escrow; (4) insurance policy; and (5) the ODR service provided by SquareTrade until June 2008, which was replaced by the eBay Dispute Resolution Centre in 2008 (Wang, 2017a, 27).

11  Until June 2008, eBay’s preferred dispute resolution provider, SquareTrade, helped eBay users who had disputes concerning eBay transactions. SquareTrade’s position was almost that of an in-house dispute resolution provider, as eBay referred its users exclusively to SquareTrade through a link on its website. There were two stages in the general operation of this system. At the first stage, SquareTrade offered eBay users a free web-based forum which allowed them to attempt to resolve their differences on their own (see eBay Dispute Resolution Overview, 2021). This is known as an ‘automated negotiation platform’ (Rabinovich-Einy, 2006, 253 and 258). When settlement could not be reached at the first stage, SquareTrade offered the use of a professional mediator for a nominal fee, with the remaining costs being incurred by eBay (see eBay Dispute Resolution Overview). This second stage is called ‘online mediation’. (Rabinovich-Einy, 2006, 253 and 259). eBay’s use of SquareTrade had the advantage of resolving misunderstandings fairly, providing a neutral go-between for buyers and sellers, reducing premature negative feedback, and generating trust in the eBay community (see eBay Dispute Resolution Overview). However, because of new developments in the eBay feedback system in May 2008, eBay and SquareTrade ceased their collaboration in dispute resolution in June 2008, though SquareTrade continues providing its warranty services and trustmark programmes to eBay users which began in 2006.

12  In 2011 Modria was established in the US as a private entity, providing a cloud based ODR platform to eBay and PayPal to resolve their online purchase disputes. It was claimed that Modria technology enabled companies to deliver fast and fair resolutions to disputes of any type and volume (see Tyler Technologies, Modria). It has been estimated that its ODR services at eBay and PayPal process 60 million cases per year (Rule, 2010). In 2017 Modria was acquired by Tyler Technologies to provide a wider service to courts and access to justice (see Tyler Technologies, Modria).

13  If consumers shop on other platforms without embedded dispute resolution services in the EU, they can try to email, phone, or live chat with sellers, or submit a claim via the EU ODR portal. In 2016 the single entry public ODR platform was established and safeguarded under Regulation (EU) 524/2013 of the European Parliament and of the Council on Online Dispute Resolution for Consumer Disputes (‘ODR Regulation’). Its objective was to harmonize ODR practice and build consumer trust in online marketplaces in the EU. It was the first available EU-wide web-based dispute resolution platform for consumers and traders to resolve their contractual disputes about online purchases of goods and services. The ODR platform enables online purchase disputes to be resolved out of court in a more affordable, simple, and timely manner. Online traders are obliged to provide a link to the ODR platform (European Commission, Online Dispute Resolution Trader’s Section). The ODR platform has several unique features compared with commercial ODR platforms on the market. The most distinctive aspects are that the process is regulated, and the standards of dispute resolution bodies are checked to ensure consistency. After submitting a complaint form via the ODR platform, consumers have 30 days to agree with traders on the dispute resolution body that will automatically receive the details of the dispute from the ODR platform (Art 9 (8) ODR Regulation). Dispute resolution bodies which are registered with the national authority and meet EU standards are listed on the EU ODR platform (Art 5 (4) ODR Regulation). In addition to this, the single entry ODR platform also has two other unique features compared with commercial ODR platforms:

• •  It allows consumers to use any of the 24 official languages of the EU;

• •  It has a national contact point to provide assistance to users at all stages (see ODR User Guide, 2018).

14  These unique features aim to provide much greater certainty and confidence to users in the EU, though there is scope for improvement of both legal and technological measures. Currently there are no embedded online hearing technologies on the ODR platform. It would be helpful to integrate advanced dispute resolution technologies into the EU ODR Portal to enhance the ODR process on the ODR platform.

15  Moreover, although it is compulsory for online sellers to provide a link to the portal on their website, it is not compulsory for online vendors to respond on the portal. As already mentioned, the EU ODR portal does not provide online hearings, nor does it provide decision-making services. It acts as a case filing platform with a list of accredited dispute resolution bodies for both sellers and consumers to agree on if they so desire.

16  If a solution cannot be reached in mediation, the buyer will still be able to submit a claim in a court under the European Small Claims Procedure—a fast track system with the judgment enforceable in all EU Member States (A Guide for Users to the European Small Claims Procedure, 2014).

17  If the online vendor is located in a non-EU state, it is less likely consumers will obtain any response from them when using the EU ODR Portal. If there is no response from the company, the case will be automatically closed after 30 days. Under these circumstances, it is advisable to check with consumer associations to see whether the company has subscribed with an ADR body, and if so, to submit a claim to the designated ADR body. ADR should here be understood to be in the form of mediation but not arbitration, as traders cannot force consumers to use arbitration and prevent them from going to court. One reason for this is the unequal bargaining power between traders and consumers.

18  If a consumer is located in London and the online trader is located in Switzerland, as neither country is an EU member state, they cannot use the EU ODR portal. However, they can still access ADR entities in EU countries. They should check with consumer associations to see whether the company has registered with an ADR body, and if so, submit a claim there. If a solution cannot be reached in mediation, UK consumers can use a court service called ‘Money Claim Online’ for a sum less than £100,000, against no more than two people or organizations with addresses in the UK (see Money Claim Online (MCOL), ‘User Guide for Claimants’). Thus, if the consumer wanted to use this service, the online trader in Switzerland must have a branch and address in the UK.

19  Consumer protection is usually enforced by national law. EU consumer legislation ensures a harmonized level of protection to EU consumers within the Union. Consumers outside the EU are subject to protection under their own national laws. At the international level, the United Nations Commission on International Trade Law (UNCITRAL) Technical Notes on ODR (2017) provide a non-binding ODR procedural reference, which is intended to foster development of ODR for use in both business-to-business (‘B2B’) and business-to-consumer (‘B2C’) disputes arising from cross-border low-value sales or service contracts concluded using electronic communications. The Technical Notes were the result of the UNCITRAL working group drafting the procedural rules on online dispute resolution for cross-border electronic commerce transactions (UNCITRAL Technical Notes on ODR). It is argued that using the Technical Notes may help alleviate the legal uncertainty of using ODR systems by promoting party autonomy and implementing core legal principles with technology neutral clauses (Wang, 2017a, 8).

20  As discussed above, although it is desirable to have a specialized and integrated ODR platform for a complete dispute resolution process, a wide range of communication application software and video conferencing facilities such as email, web-based forms, videoconference, or videotelephony can be used to assist in resolving disputes online with vendors directly, or via dispute resolution bodies or small claim courts. Parties involved in dispute or ADR bodies can choose the most suitable software with their desired functions as the platform to conduct hearings.

21  Specialized ODR software packages which can be tailored to suit the needs of users are available. Examples of specialized ODR software without embedded AI technologies, with case management tools and auditing facilities include Decider, a UK company which provides ODR services and offers commercial software with secure internal messaging, case management tools, and auditing facilities (see Decider by Resolver Group). Caseload Manager, a US company, offers cloud-based commercial software and a subscription model based on the number of new cases annually (see Caseload Manager). Although these specialized ODR software packages are not embedded with AI technologies, additional AI-powered tools such as Fireflies, Microsoft Teams, and Google Live Transcribe can be used to transcribe live audio into text to assist evidence gathering and analysis in the ODR process (see Fireflies, Microsoft Teams, and Android Live Transcribe websites).

22  In recent years legal practice and services have been embedded with increasingly sophisticated digital technologies (Frostestad Kuehl, 2019) including encrypted data and information, blockchain, service-oriented computing, cloud computing, artificial intelligence, social media, and electronic forensic evidence. Service-oriented computing ‘offers a promising solution in discovering other appropriate agents, reaching agreements between service providers and customers, managing the joint execution of tasks and dealing with any problems that arise’ (Wang and Griffiths, 2010, 156). The more recent blockchain technology—a shared and distributed ledger and an alternative to traditional databases—supports the automated execution of smart contracts (Daniel and Guida, 2019, 46–53). It is argued that blockchain and service-oriented computing may be combined as ‘service-oriented permissioned blockchain’ to optimize services such as verifiable data (Qiu and others, 2020, 203–15). On a blockchain, all parties may remain anonymous, arbitrators may implement decisions using private key(s), and an arbitral award may be self-executed in an ‘automatic dispute resolution process’ (LawTech UK/UK Jurisdiction Taskforce Digital Dispute Resolution Rules, 2021, 13). That is, an arbitrator may be given the technical power to operate, modify, sign, or cancel any relevant digital asset if the arbitrator has been given, eg, the necessary private key(s). From a technological point of view, it would be worrying if an arbitrator could be given a party’s private key(s) to manage digital assets. There is need to look further into the features of the special category of disputes concerning digital assets, and to work on necessary guidelines concerning the security and due process of arbitrators’ implementation of decisions directly on-chain in an automated dispute resolution process. An automatic dispute resolution process means:

These advanced technologies may be broadly employed in the ODR process to execute arbitration awards or mediation agreements, and shape online written communication to avoid escalating conflict by blocking inappropriate language in the process (Wang, 2017a, 80). It is also argued that blockchain technologies may improve the credibility and authenticity of electronic evidence in judicial services (Lu, 2021, 3). In Hangzhou Huatai Media v Shenzhen Daotong Technology, Hangzhou Internet Court of the People’s Republic of China ruled that the electronic evidence data stored in the blockchain has a clear source, its generation and transmission path is definite and clear, and it can be mutually verified with the screenshots of web pages, source code information, and call logs. Therefore, the generated electronic data is reliable (Hangzhou Huatai Media Culture Media v Shenzhen Daotong Technology Development, 2018; English summary provided by the Supreme People’s Court of China, 12 April 2019).

23  In the partly AI-assisted ODR environment, AI technology may help, as mentioned earlier, in transcribing evidential audio to text, and provide a provisional outcome for a human mediator or arbitrator to review and make the final decision. In the entirely AI-enabled ODR service environment, once a complaint is filed and accepted, and relevant evidence is submitted, a robotic mediator or arbitrator may help one to resolve the issue without any human intervention. In the medium to long term, a dispute resolution intelligence machine itself may be able to collect and analyse critical data concerning previous decision-making by arbitrators or mediators and understand the nature of the dispute and the associated rules and laws relating to the case (Wang, 2017a, 98). In 2015 ‘Arbitrator Intelligence’, a global information aggregator, was set up to collect arbitrators’ past decision-making information including published and unpublished international arbitral awards (Arbitrator Intelligence).

24  Trust is the most crucial factor in building a successful ODR. It has been suggested that there are three fundamental factors for a successful ODR system: ‘trust, convenience and expertise’, and it is important to strike the balance among these factors depending on the intended use of the ODR system in question (Katsh and Rifkin, 2001, 73). The design of the system will depend on who the parties are and what the context is (Katsh and Rifkin, 2001, 75). There is often a trade-off between the power of an application—expertise—and how complicated it is to use—convenience—(Katsh and Rifkin, 2001, 76). There is a relationship between expertise and trust, as a seemingly competent system is more likely to be trustworthy. Consequently, convenience and trust may be inversely related. With the deployment of advanced ODR systems embedded with AI and other robotic technologies such as machine learning, the fundamental factors of trust, convenience and expertise may need to be expanded to a wider scope of factors such as: authenticity, reliability, durability, efficiency, and flexibility. In both basic and advanced/intelligent ODR systems, trust is of paramount importance, in particular in advanced ODR systems where a higher level of trust is desired due to the fact that users are less willing to trust computer algorithms to make automated decisions (Wang, 2017a, 86–87).

25  It was suggested that ‘Intelligent machines hold the promise of more rational, consistent, and unbiased decisions when compared to human actors’ (Eidenmüller, 2019, 114). However, that is not necessarily true as it is challenging for AI algorithms to understand context at the current stage of technological advancement. It would be almost impossible to block all inappropriate language as a statement may only be able to be understood correctly within the context. Moreover, it was pointed out that the ability of the machine as an algorithmic actor is currently very limited as to ‘identify the issues and legal framework, establish facts, evaluate the parties’ interests, disaggregate issues, establish positions, exchange information, suggest options for resolution, set out a time frame for actions, seek agreement and create binding resolutions’ (Morison and Harkens, 2019, 618–35).

26  If AI technology is employed for negotiation, it is generally feasible for parties to use AI technology to assist their negotiation, as long as parties trust the algorithms and can therefore accept the outcome. For example, a negotiation starts when a player represented by AI technology introduces a statement that he/she wants to justify. A negotiation ends if the opponent represented by AI technology accepts the statement—justified—or if the statement is withdrawn—not justified—which is so called the ‘DiaLaw’ game (Lodder and Zeleznikow, 2005, 305).

27  If AI technology is employed for mediation, legal feasibility may be challenged. In a partly AI-assisted mediation process, a human mediator can review the case and make the final decision on the mediation agreement. Such mediation agreement will be valid if parties agree and sign it. In an entirely AI-enabled mediation environment, an automated mediation settlement may be further challenged if the competency and legality of robotic mediators are called into question.

28  If AI technology is employed for arbitration, legal feasibility may be hotly debated. It has been suggested that ‘the existing legal framework for international commercial arbitrations, the “New York Convention” (NYC) in particular, is capable of adapting to and accommodating fully AI-powered arbitrations’ (Eidenmüller and Varesis, 2020, 5). However, although arbitration improves the chance of enforceability of the outcome in another country due to the established Convention on the Recognition and Enforcement of Foreign Arbitral Awards (1958) (‘New York Convention’), moving arbitration online, either partly AI-assisted or entirely AI-enabled, has challenged due process in arbitration procedures and the enforceability of arbitral awards. For example, the requirements in Article II (2) New York Convention that ‘[t]he term “agreement in writing” shall include an arbitral clause in a contract or an arbitration agreement, signed by the parties or contained in an exchange of letters or telegrams’ will be challenged. In July 2006, UNCITRAL recommended that Article II (2) should be interpreted widely ‘recognizing that the circumstances described therein are not exhaustive’ (Recommendation regarding the Interpretation of Article II, Paragraph 2, and Article VII, Paragraph 1, of the Convention on the Recognition and Enforcement of Foreign Arbitral Awards, 2006, 62) However, it is still uncertain whether the interpretation will be extended to arbitration agreements and awards which are machine-generated—or AI-enabled—without human intervention. In particular if the governing law of the main contract did not match the seat of arbitration, it challenged applicable law governing an arbitration agreement for enforcement (Kabab-Ji SAL (Lebanon) v Kout Food Group (Kuwait), 2020). If the AI-enabled seat of arbitration did not match the chosen law for the main contract, it would be further challenged because the intention of parties is unknown by computing codes. The content of a smart contract may be incompatible with the parties’ mutual intent (Werbach and Cornell, 2017, 318).

29  Whether AI-assisted and AI-enabled negotiation, mediation, and arbitration is legally feasible depends on the interpretation of current and future relevant regulations. For example, the Proposal for AI Regulation prohibited certain AI practices such as an AI system ‘materially distorting a person’s behaviour’ which may cause ‘physical or psychological harm’, or lead to ‘detrimental or unfavourable treatment of certain natural persons or whole groups’ (Art 5 Proposal for AI Regulation). Therefore, an AI ODR system in the EU needs to be assessed by risk factors and avoid prohibited AI practices. However, both a harmonized risk assessment framework and accredited bodies in countries and regions need to be established.

30  In a partly AI-assisted arbitration process a human arbitrator can review the case and make the final decision, so the current general rules concerning the competence of arbitrators can still apply but arbitrators’ competence may need to be extended to include digital competence. Parties may form an arbitration agreement by electronic means, hearings can be conducted online, and arbitrators may issue an arbitral award by electronic means (Wang, 2017a, 107–47). The validity of electronic arbitration agreements and arbitral awards may require further interpretation and justifications if national law requires them to be in writing (Wang, 2017a, 110–29). During online live hearings, due process can also be further challenged, for example whether online hearings ensure the same level of accountability, impartiality, transparency, and independence as the in-person hearing process, where parties sit in the same room. Another issue is that an AI-assisted online hearing system should avoid ‘algorithmic bias and discrimination’, for which US legislative instruments have been proposed to tackle the issue of algorithmic accountability (Chae, 2020, 18).

31  In an entirely AI-enabled arbitration environment, due process will be further challenged. The competency and legality of robotic arbitrators may be of great concern. Ethical considerations may also emerge when contrasting traditional arbitrators, and web-based visible or invisible robot arbitrators. The practicality of ensuring fairness and impartiality in robotic arbitration will affect how feasible an entirely AI-enabled robotic arbitration is in the future. For example, the professional background of a traditional arbitrator is transparent to the parties, allowing a relationship of trust to form. However, knowing what a robot arbitrator is programmed with is a completely different question and it may be difficult for parties to trust that there is no bias in the decision-making process.

32  New ethical principles such as accuracy of language from machine learning may need to be developed. For example, if service providers use AI technologies for natural language processing, national language processes will be able to recognize speech and transcribe the parties’ speeches for evidence during the hearings. National language processing can also provide understanding of different languages and provide an automated response to questions. However, whether AI technology can understand the legal issues in context is of great concern. It has been noted that for ‘an artificial intelligence model to replace the arbitrator, the machine itself must understand the nature of the dispute and the associated rules and laws relating to the case’ (Wang, 2017a, 98). The AI model may need to have knowledge of previous cases and understand how any differences between previous cases and current cases may affect the relevance to the current cases in question (Main and others, 2001, 1–28). In order to achieve this, relevant technological measures such as machine learning and expert systems should be utilized. It has been suggested that the task domain of a robotic arbitrator may be represented using a variety of high-level technologies including service-oriented computing, and automated computing, joining up with lower-level technologies such as expert systems, machine learning, argumentation theories, systems of non-monotonic logic, case-based reasoning, and knowledge-based method (Wang, 2017a, 99).

33  Moreover the quality of data is of great essence to the quality of ODR service. Original and raw data may be checked and trained by human experts to improve the accuracy—known as ‘trained data’ or ‘training data’. The trained data model is ‘the algorithm with the set of parameters that optimized performance on the training dataset’ and is then being ‘put to work on a new test dataset, to see how effective it is at predicting outside the original training sample’ (Eidenmüller and Varesis, 2020, 7).

34  As discussed above, even if the process of an entirely AI-enabled arbitration is technologically feasible, there is still uncertainty over the validity of automated arbitration agreements by clause builder algorithms and the enforceability of automated arbitral awards without the review of a natural person/expert. This is an area for further consideration for legal reform.

35  In the last two decades ODR to resolve disputes online has developed significantly. Although online arbitration may be more cost effective than traditional arbitration, it challenges traditional national and international arbitration laws such as the New York Convention and requires specific interpretation. The ODR Regulation offers procedural rules to use the EU ODR portal to resolve consumer disputes but does not offer online arbitration or mediation. It simply acts as a case filing system rather than an interactive dispute resolution platform embedded with hearing facilities. Even though there is no specific legislation concerning online arbitration for commercial disputes, there are self-regulatory developments such as a wide range of online arbitration and mediation procedures or rules by arbitration organizations worldwide, and guidance notes from international organizations such as UNCITRAL’s Technical Notes on ODR.

36  There are no simple answers or solutions as to how and when fully AI-enabled arbitration can be wholly accepted as accessible, reliable, efficient, and cost-effective private justice, and whether private judicial systems can form a partnership with public judicial systems to ensure enforceability. However, it is clear that it is unavoidable that modern-day society is heading in the direction of automation in all aspects of our lives including judicial services. Legal practitioners, arbitrators, mediators, and lawyers should develop their digital competence in the age of artificial intelligence. Ethical, legal, and technological standards for using AI-enabled dispute resolution should be harmonized internationally to improve legal certainty and access to justice.

Faye Fangfei Wang

Cited Bibliography

• EM Katsh and J Rifkin, Online Dispute Resolution: Resolving Conflicts in Cyberspace (Jossey-Bass San Francisco 2001).

• J Main, T Dillon, and S Shiu ‘A Tutorial on Case Based Reasoning’ in SK Pal, TS Dillon, and DS Yeung (eds), Soft Computing in Case Based Reasoning (Springer London 2001) 1–28.

• RS Haydock and JD Henderson, ‘Arbitration and Judicial Civil Justice: An American Historical Review and a Proposal for a Private/Arbitral and Public/Judicial Partnership’ (2002) 2(2) Pepperdine Dispute Resolution Law Journal 141–98.

• AR Lodder and J Zeleznikow, ‘Developing an Online Dispute Resolution Environment: Dialogue Tools and Negotiation Support Systems in a Three-Step Model’ (2005) 10 Harvard Negotiation Law Review 297–337.

• O Rabinovich-Einy, ‘Technology’s Impact: The Quest for a New Paradigm for Accountability in Mediation’ (2006) 11 Harvard Negotiation Law Review 253–93.

• D Miller-Moore and M Jennings, ‘At the Forefront of ODR: Recent Developments at the AAA’ (2007) 62(1) Dispute Resolution Journal 5.

• F Wang, Online Dispute Resolution (ODR): Technology, Management and Legal Practice from an International Perspective (Chandos/Elsevier Publishing Oxford 2009).

• M Taddeo, ‘Defining Trust and E-Trust: From Old Theories to New Problems’ (2009) 5(2) International Journal of Technology and Human Interaction 23–35.

• F Vicuña, ‘Mediation’ in R Wolfrum (ed), The Max Planck Encyclopedia of Public International Law (OUP Oxford 2010), <www.mpepil.com> (accessed 1 October 2021).

• C Rule, ‘Using Technology to Manage High Volume Caseloads: The eBay/PayPal Experience’ (1 November 2010) <https://www.archives.gov/files/ogis/events-presentations/acus-colin.pdf> (accessed 1 October 2021).

• F Wang and N Griffiths, ‘Protecting Privacy in Automated Transaction Systems: A Legal and Technological Perspective in the European Union’ (2010) 24(2) International Review of Law, Computers & Technology 153–62.

• K Hakapää, ‘Negotiation’ in R Wolfrum (ed), The Max Planck Encyclopedia of Public International Law (OUP Oxford 2013), <www.mpepil.com> (accessed 1 October 2021).

• J Kaplan, Artificial Intelligence: What Everyone Needs to Know (OUP Oxford 2016).

• F Wang, Online Arbitration (Informa Law from Routledge Oxford 2017a).

• F Wang, ‘Online Dispute Resolution: Best Practices in Comparative Perspective’ (2017b) 12(2) Journal of Comparative Law 472–95.

• K Werbach and N Cornell, ‘Contracts Ex Machina’ (2017) 67 DukeLJ 313–82, <https://scholarship.law.duke.edu/dlj/vol67/iss2/2> (accessed 1 October 2021).

• C Menkel-Meadow, ‘Alternative Dispute Resolution’ in H Ruiz Fabri (ed), The Max Planck Encyclopedia of International Procedural Law (OUP Oxford 2018), <www.mpeipro.com> (accessed 1 October 2021).

• F Daniel and L Guida, ‘A Service-Oriented Perspective on Blockchain Smart Contracts’ (2019) 23(1) IEEE Internet Computing 46–53.

• H Frostestad Kuehl, ‘Technologically Competent: Ethical Practice for 21st Century Lawyering‘ (2019) 10 Case Western Reserve Journal of Law, Technology & the Internet 1–30.

• H Eidenmüller, ‘Machine Performance and Human Failure: How Shall We Regulate Autonomous Machines?’ (2019) 15 Journal of Business & Technology Law 109–14.

• J Morison and A Harkens, ‘Re-Engineering Justice? Robot Judges, Computerised Courts and (Semi) Automated Legal Decision-Making‘ (2019) 39(4) Legal Studies 618–35.

• R Thümmel, ‘International Centre for Dispute Resolution: American Arbitration Association (AAA)’ in H Ruiz Fabri (ed), The Max Planck Encyclopedia of International Procedural Law (OUP Oxford 2019), <www.mpeipro.com> (accessed 1 October 2021).

• Y Chae, ‘US AI Regulation Guide: Legislative Overview and Practical Considerations’ (2020) 3(1) Journal of Robotics, Artificial Intelligence & Law 17–40.

• H Eidenmüller and F Varesis, ‘What is an Arbitration? Artificial Intelligence and the Vanishing Human Arbitrator’ (17 June 2020), <https://ssrn.com/abstract=3629145> (accessed 1 October 2021).

• C Qiu, H Yao, FR Yu, C Jiang, and S Guo, ‘A Service-Oriented Permissioned Blockchain for the Internet of Things’ (2020) 13(2) IEEE Transactions on Services Computing 203–15.

• J Chaisse and Q Xu, ‘Conservative Innovation: The Ambiguities of the China International Commercial Court’ (2021) 115 AJIL Unbound 17–21.

• T Lu, ‘The Implementation of Blockchain Technologies in Chinese Courts’ (4 January 2021) Stanford Journal of Blockchain Law & Policy, <https://stanford-jblp.pubpub.org/pub/blockchain-in-chinese-courts> (accessed 1 October 2021).

Cited Documents

• AAA Mediation.org, ‘Mediation Procedures’ (1 August 2013).

• ACICA, ‘Draft Procedural Order for the Use of Online Dispute Resolution Technologies’ (2016) <https://acica.org.au/wp-content/uploads/2016/08/ACICA-online-ADR-procedural-order.pdf> (accessed 21 April 2022).

• ACICA, ‘Online Arbitration Guidance Note’ (2020) <https://acica.org.au/wp-content/uploads/2020/05/ACICA-Online-Arbitration-Guidance-Note.pdf> (accessed 21 April 2022).

• CIETAC, ‘Online Arbitration Rules’ (4 November 2014).

• Civil Justice Council ODR Advisory Group, ‘Online Dispute Resolution for Low Value Civil Claims’ (February 2015).

• Convention on the Recognition and Enforcement of Foreign Arbitral Awards (done 10 June 1958, entered into force 7 June 1959) 330 UNTS 38 <https://uncitral.un.org/sites/uncitral.un.org/files/media-documents/uncitral/en/new-york-convention-e.pdf> (accessed 21 April 2022) (New York Convention).

• European Commission, ‘A Guide for Users to the European Small Claims Procedure’ (2014) <https://e-justice.europa.eu/fileDownload.do?id=1561ac8a-41fd-482c-ba42-7ed16180d541> (accessed 1 October 2021).

• European Commission, ‘ODR User Guide’ (25 October 2018) <https://ec.europa.eu/consumers/odr/userguide/pdf/MA/userguide-en.pdf> (accessed 1 October 2021).

• Hague Conference on Private International Law, ‘Convention on Choice of Court Agreements’ (concluded 30 June 2005, entered into force 1 October 2015) (2005) 44 ILM 1294 <https://www.hcch.net/en/instruments/conventions/full-text/?cid=98> (accessed 21 April 2022) (Hague Convention on Choice of Court Agreements).

• LawTech UK/UK Jurisdiction Taskforce, ‘Digital Dispute Resolution Rules’ (2021).

• LCIA, ‘Updates to the LCIA Arbitration Rules and the LCIA Mediation Rules’ (2020) <https://www.lcia.org/lcia-rules-update-2020.aspx> (accessed 1 October 2021).

• Proposal for a Regulation laying down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act) and amending certain Union Legislative Acts (21 April 2021) COM(2021) 206 final, 2021/0106 (COD).

• Recommendation regarding the Interpretation of Article II, Paragraph 2, and Article VII, Paragraph 1, of the Convention on the Recognition and Enforcement of Foreign Arbitral Awards, Done in New York, 10 June 1958 (adopted 7 July 2006) UN Doc A/61/17, Annex II.

• Regulation (EU) 524/2013 of the European Parliament and of the Council on Online Dispute Resolution for Consumer Disputes and amending Regulation (EC) No 2006/2004 and Directive 2009/22/EC (21 May 2013) [2013] OJ L165/1.

• UNCITRAL, ‘Technical Notes on Online Dispute Resolution’ (2017).

Cited Websites and Services

• AAA, ‘Virtual Hearing Managed Services’ <https://adr.org/virtual-hearing-managed-services> (accessed 1 October 2021).

• ACICA, ‘Practice & Procedures Toolkit’ <https://acica.org.au/acica-practice-procedures-toolkit/> (accessed 1 October 2021).

• American Arbitration Association and International Centre for Dispute Resolution, ‘Virtual Hearings’ <https://go.adr.org/covid-19-virtual-hearings.html> (accessed 16 May 2022).

• Android Live Transcribe <https://www.android.com/accessibility/live-transcribe/> (accessed 1 October 2021).

• Arbitrator Intelligence, ‘FAQs’ <https://arbitratorintelligence.com/faqs/> (accessed 1 October 2021).

• Canadian Civil Resolution Tribunal <https://www.civilresolutionbc.ca/> (accessed 1 October 2021).

• Caseload Manager <https://www.caseloadmanager.com/> (accessed 1 October 2021).

• CEDR, ‘Telephone and Online Meditations’ <https://www.cedr.com/commercial/telephone-and-online-mediations/> (accessed 1 October 2021).

• China International Commercial Court, ‘One-Stop Dispute Resolution Centre with Diversified Settlement Mechanisms for International Commercial Disputes’ <http://cicc.court.gov.cn/html/1/218/321/323/index.html> (accessed 1 October 2021).

• Decider by Resolver Group <https://www.resolvergroup.com/in/careers-at-resolver/www.resolvergroup.com> (accessed 1 October 2021).

• eBay, ‘Dispute Resolution Overview’ (2021) <http://pages.ebay.com/services/buyandsell/disputeres.html> (accessed 1 October 2021).

• eBay, ‘SquareTrade Warranties on eBay’ (2021) <https://pages.ebay.co.uk/safetycentre/buyingwithconfidence/warranties/index.html> (accessed 1 October 2021).

• European Commission, ‘Online Dispute Resolution Trader’s Section’ <https://ec.europa.eu/consumers/odr/main/?event=main.trader.register> (accessed 1 October 2021).

• European Commission, ‘Resolve Your Consumer Complaint’ <https://ec.europa.eu/info/live-work-travel-eu/consumer-rights-and-complaints/resolve-your-consumer-complaint_en> (accessed 1 October 2021).

• Fireflies <https://fireflies.ai/product/features> (accessed 1 October 2021).

• GOV.UK, ‘Make a Money Claim Online’ <https://www.gov.uk/make-money-claim> (accessed 1 October 2021).

• HiiL, ‘Rechtwijzer 2.0’ (2021) <www.hiil.org/project/rechtwijzer> (accessed 30 April 2021).

• HiiL, ‘The Technology of Access to Justice: Rechtwijzer 2.0’ (2021) <https://www.hiil.org/news/the-technology-of-access-to-justice-rechtwijzer-2-0/> (accessed 1 October 2021).

• HM Courts & Tribunals Service, ‘Money Claim Online (MCOL)—User Guide for Claimants’ <https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/458835/money-claim-online-user-guide.pdf> (accessed 1 October 2021).

• Microsoft Teams: Online Meetings <https://www.microsoft.com/en-us/microsoft-teams/online-meetings> (accessed 1 October 2021).

• Traffic Penalty Tribunal <https://www.trafficpenaltytribunal.gov.uk/want-to-appeal/> (accessed 1 October 2021).

• Tyler Technologies, ‘Modria’ <https://www.tylertech.com/products/modria> (accessed 1 October 2021).

• Youstice <https://web.archive.org/web/20210127064431/https://www.youstice.com/en/> (accessed 16 March 2022).

Cited Cases

• Halsey v Milton Keynes General NHS Trust, England and Wales Court of Appeal (11 May 2004) [2004] EWCA Civ 576.

• Hangzhou Huatai Media Culture Media Company Limited v Shenzhen Daotong Technology Development Company Limited, First Court No 81, Hangzhou Internet Court of the People’s Republic of China (27 June 2018) Zhe 0192 Civil Case <http://english.court.gov.cn/2019-12/04/content_37527759.htm> (accessed 1 October 2021).

• Kabab-Ji SAL (Lebanon) v Kout Food Group (Kuwait), England and Wales Court of Appeal (20 January 2020) [2020] EWCA Civ 6.