Artificial Intelligence has recently proven to play a significant role in transforming the processes of dispute resolution. A prime example is eBay Resolution Centre which eBay designs based on the patterns they identified on dispute emergence between buyers and sellers and how they are resolved. The whole system is questionnaire-based which helps in identifying the key factors leading to disputes and offers the best possible resolution options, thus playing the role of a mediator as well. This indeed has proven to be a viable option as hiring expert individuals to facilitate the ODR process with tens of thousands of cases piling up daily, with most of them being very low valued is just not efficient or cost-effective. 

Today, eBay Resolution Center automatically resolves 90% of its caseload without the involvement of a human third party. 

The debate regarding the degree of autonomy possessed by software while providing dispute resolution services boils down to the point of whether AIs are capable of replacing human-centered dispute resolution involving all the processes of mediation and arbitration. While the examples stated and the flow of technological innovations point to the obvious answer, one thing to be kept in mind is that the concept of Alternate Dispute Resolution arises

from the preferences of its stakeholders i.e disputed parties and their choice of not wanting to take the exhaustive court methods of case settlements and coming to a peaceful consensus that benefits all parties involved. Therefore, the choice of opting for AI automated systems as a third party in ODR largely depends on the preferences and experiences of the stakeholders (disputed parties). 

The ‘process’ (including negotiation, mediation, arbitration, and method of delivery of factual information) by which dispute resolution takes using technologies is extremely important in determining the procedural justice experience of disputing parties. Whether the process involves automated systems at the beginning of dispute resolution; if automation is just used in collecting the factual information and then passed on to a human mediator/arbitrator thus AI only playing the role of a fourth party; or using AI-based systems throughout the whole process of dispute resolution. Each of these routes can affect not just the procedural justice experience of stakeholders involved but also to an extent stimulate the outcome reached in the final case settlement process. 

Hence the role of AI in stimulating the case settlement outcomes can be mainly narrowed down to certain perceived aspects such as: 

1. The level of communication made through AI – AI can offer very different modes of communication wherein usually cases involving a mere exchange of information, facts & figures regarding eCommerce complaints do not require any high level of interpretation or 

comprehensions and normal textual language-based system. However, in turn for cases such as child custody, divorce, or property settlement, mere textual systems do not do the job. 

2. Binding or non-binding nature of the outcome- The nature of the judgment made during the dispute resolution process can also affect the acceptance of AI in the process. If the judgment is binding, not all parties would be comfortable in letting a machine judge the outcome of the underlying dispute. The case may be different if the agreement is non-binding. This again depends on the faith the general public develops in the accuracy of judgments made by AI systems. 

3. Absence of non-verbal cues- Non-verbal cues such as gestures, facial expressions, voice connotations, etc. help the listeners make sense of the textual words and interpret their

meaning. Non-verbal cues are especially important in dispute resolution as they can lead to unfavorable circumstances where the disputing parties can miscommunicate the intentions and fill in the communication gap according to their ability to comprehend them. This can  often increase the chance of creating more conflicts than resolving them. AI-based systems can give clear and logical judgments based on the past history of judgments on certain cases or data input from the existing laws and clauses however they have limited capabilities in grasping the non-verbal cues of human beings. Although there are AI-based Facial recognition systems that also can identify certain emotions like happiness, anger, sadness, etc. to identify the state of emotion of subjects; combined with machine learning, AI can be used to identify non-verbal cues to an extent. However, the accuracy is not 100% proven. So the question again arises whether the procedural justice experience can be fully entrusted to AI? 

4. Notion against ‘machine-made justice’- This again refers to how much people can trust AI systems, essentially machines to make accurate judgments not just based on the previous history of recorded cases or data inputs of laws and regulations but identifying the key aspects of every new case; including evidence collection, analysis, verification of facts, etc. and making the correct judgments based on predictions. Such 'Machine-made Justice' is usually not trusted enough to provide better results than the one made by humans. After all, machines are the products of human minds. 

So the acceptance of AI into dispute resolution boils down to the debate of Human VS AI wherein the complete involvement of AI as a third party mediator/arbitrator pokes around several questions on whether it affects the quality of the process as well as general acceptability. Complicated cases regarding real estate properties, divorce, and child custody cannot be concluded on certain streamlined guidelines already programmed into AI, rather they require interpretation and comprehension regarding all situations and stakes at the hand of parties involved which certainly cannot be handed down completely to an automated machine system.