ARTIFICIAL INTELLIGENCE IN MEDIATION
This paper aims to identify the use of Artificial Intelligence in Mediation and the benefits it may have as compared to mediation in the conventional sense. Artificial Intelligence is a developing concept and will benefit almost all fields of life. Thus, mediation is no different when we say Artificial Intelligence will bring about a more seamless process. Further, this paper will look into the various existing software developed to benefit people opting for mediation. This paper will also look into any new areas where such software may further benefit the process. Lastly, this paper will look into the impact other jurisdictions have seen after implementing Artificial Intelligence in the process of mediation.
With the development of technology, it is hard to find any process that does not use any form of technology. The benefits one may derive from reducing human interference have been made abundantly clear over the last few decades. In recent times the development of Artificial Intelligence (hereinafter “AI”)and machine learning have taken technology a step further.
We must first understand what AI is? AI refers to the teaching of human intelligence to machines programmed to perform and simulate the functions of humans. The term can be applied to any machine that learns psychological characteristics, such as learning and problem-solving. The ideal feature of AI is the ability to rationalize and take on tasks that can be very effective in achieving the goal of choice.[i]
Machine learning teaches computers, through programmed algorithms, to perform human actions and allows the program to gather knowledge, from previous use, to develop their actions further. This is a process whereby a computer collects data, through repeated use, and uses it to improve and expand the functioning of its programs, to create life–like observations and functioning. Such a program does not need to be constantly developed. Instead, the machine learns by itself through a collection of data from previous interactions.[ii]
The next concept we must look into is mediation. Mediation is a third-party-driven negotiation where there is a neutral third party that helps resolve a dispute. This is a non-binding process that falls under the category of alternate dispute resolution or out-of-court settlement. The benefits of mediation are that it is a confidential process that is preferred by companies and even in family disputes. This process looks into the merits of a case and the interests of each party in the dispute.[iii]
In this paper I shall first analyse the relation between AI and Mediation. Subsequently, I will look into some AI programs that already exist in the field of mediation and how they function. Further, I shall look into some other areas where AI may be used to further simplify the process of mediation. Lastly, I shall delve into the benefits and problems that AI may have, while being used for the process of mediation
III. The interrelation between Artificial Intelligence and Mediation
AI is still a developing concept, and thus there is not much practical use of it yet in the process of mediation. Mediation has many components where AI will reduce the human element. The main aim of mediation is dispute settlement, and through AI, this can be made easier and quicker. Initially, online mediation included only communication through emails and a platform for video – conferencing. This was more or less the same as in-person mediation and did not involve any significant change to the functions of the mediator or to the duties of either party. But over time, this has changed. There have been developments where software has begun taking over certain functions of the mediator. Some software has been developed to help the parties identify the strengths and weaknesses of their case and understand a possible solution.
IV. Existing programs in mediation
There have been many noteworthy programs in mediation. They have helped settle disputes in many ways and simplified the process. Some are:
SmartSettle is an advanced system that can support many parties that have goals that differ from each other or are in conflict. It can resolve high-level disputes that parties may have. SmartSettle acts as a qualified mediator for parties that may not be familiar with the process. It uses mathematical formulae to improve the dispute’s outcome by taking into consideration the various priorities a party may have.[iv]
Each Party must enter the various considerations it has. This is not disclosed to the other party and thus keeps the same level of confidentiality as the traditional mediation process has. The considerations may be either position, which is the stand a party takes on a particular issue, or the interests each party may have, which is the reasoning behind holding such a position. The process also requires parties to enter the optimal outcome for them. The parties must again enter their acceptable range for agreement on each issue known as the Zone of Possible Agreement.[v] SmartSettle has many ways at which an agreement may be arrived. The traditional method, where offers are exchanged and counter–offers are made to reach a solution. They may also have the software, consider all the parties’ entries and come up with the optimal solution. The program also caters to a situation where the parties may reach a roadblock. The program may be asked to resolve it, and it will put forth a proposal it thinks would be beneficial to both parties. Lastly, the software in the final stage of the mediation may also make suggestions that bring about a further optimized solution which the parties may not have considered. If the parties are satisfied with the same, they may end the process and consider the settlement a success.[vi]
ii. Family Winner
In Family Winner, the program uses both game theory and heuristics. This program presents essential values to show the amount to which each party wishes to be awarded the issue being considered. The program takes information from the parties and uses it to make trade-off rules, which are then used to allot issues according to the logrolling strategy. Family Winner then uses the information, the parties have entered, to form trade-off values to show what result the allocation of an unallocated issue would have.[vii]
The software basically makes suggestions based on the information the users have entered. It provides a solution that it considers would be amicable to each party. If the parties are satisfied with the same, they may move on to the next issue, which would follow the same process. But in the situation that they are not satisfied with the solution the system prompted them with, they may either make adjustments manually or, on the other hand, ask the system to provide them with an alternate solution. It also looks into which party compromised when and how many issues each party has been given in its favor. It then adjusts the remaining matters to provide an amicable solution where no party loses more than they can afford to. This software is mainly for family disputes.[viii]
iii. Automatic translation
This is a relatively mainstream program that has existed for a few years now. But the use of it in the process of mediation and international mediation specifically is significant. In the traditional mediation process, the parties would have to employ a translator to be able to have a conversation.[ix]Most business agreements and other contracts are between parties of different nationalities, and there may be a language barrier. The cost would also significantly increase when such parties are forced to employ a translator for all disputes. Such software would translate the words and read them out to parties and thus reduce the extra cost of employing a translator.[x]
iv. Automatic summarization
This program would take up the role of a mediator in summarizing. The program may record and convert the speech to text and provide a comprehensive report of all the conversations had between the mediator and the parties.[xi] This way, the mediator would be able to focus more on the conversation rather than summarizing and noting down the conversations the parties are having. This program may also provide a summary in real-time to the mediator to further enable him to understand each party’s dialogue and interests.12
V. Areas Where Artificial Intelligence can be Developed to Simplify the Process of Mediation Further
There have been some significant developments in AI to benefit the process of mediation, but there can be some further developments to benefit parties. Such developments include,
- Software to provide a Zone of Possible Agreement to the mediator. This may not be a completely AI-based program but would benefit the mediator in understanding each party’s interests and the top and bottom line. Such a program would be beneficial as it would give the mediator a clear understanding of what each party is looking for and how each party may react to a specific offer.
- A software to SMART test an offer made by the parties. SMART testing is when an offer is tested for being S – Specific, M – Measurable, A – Achievable, R – Realistic, T – Time Bound. The parties would be required to test the offer against this and put in the information against each of the heads. This would benefit the other party as well and the mediator in understanding the offer and the viability of such an offer.
- A software to exchange documents and check the originality of the same. There already exists software that allows the same, but this could be integrated into existing software or a video conferencing platform to simplify the process further. Parties would immediately be given access to the document and may view it during the session.
- Lastly, the software allows parties to identify their issues, positions, and interests and provide the same to the mediator in a user-friendly format. This would enable the parties to provide a clear understanding to the mediator. The parties may also be permitted to change the same over the course of the mediation proceedings as they are bound to change. This would greatly benefit the mediator as it would save him the time of exploring the same with the parties.
VI. Benefits And Disadvantages Of Using Artificial Intelligence in The Process of Mediation And The Impact Seen in Other Jurisdiction A. Benefits of using AI in mediation:
- No subconscious control and mental bias: a human mediator is, after all, human and would have emotions and biases ingrained in his mind. No matter how well trained a mediator, this is one setback that cannot be removed. The mediator may try to be unbiased and may achieve it to an extent in some cases, but his upbringing and values play a significant role in his thought process. AI would remove the human element entirely and thus provide an unbiased and transparent solution.[xii]
- Time – Efficient: the mediation in the traditional sense may take months on end. Though it is much faster than litigation, it still requires the parties to take time out and settle the dispute. But with the introduction of AI in mediation, the parties will be able to resolve the conflict faster and not be required to take weeks on end out from their already busy schedules.[xiii]
- Removing the chance of human error: in the traditional procedure of mediation, the parties may make an error in expressing their interests and maybe disadvantaged by the same. But with the development of AI, the parties will have a chance to view what they have written to enter into the software and correct any errors that they may have made.15
- Cost – Efficient: traditional mediation often costs a lot of money, as the mediator needs to be paid his fee, lawyers are involved, and they too need to be paid, and the parties usually meet in a neutral place, and those costs too play a role in increasing the cost of the process. But through AI, those costs are reduced by more than half. The parties usually need not meet in persona and may use a video–conferencing method to do the same. Further, the mediator’s costs are not as high as they would be when the parties have to meet in person. Using software that would almost completely replace the mediator would reduce his fee from the total cost as well. The program costs may be added, but they would not be as high as the mediator’s costs.[xiv]
- There is also no requirement for the parties to meet in person and may use video– conferencing mediums to meet and settle the dispute. If the use of a complete AI-driven mediation, the parties need not even meet through such a medium, and the entire process would seamlessly happen.[xv]
B. Disadvantages of the use of AI in mediation:
- The development may take time: developing and teaching the AI software is a time consuming process and may take many years to develop. Further, it may take even longer to fix any errors, and until they have been finetuned, the parties may be disadvantaged, due to the failure of the software, to provide an optimal result.[xvi]
- Lack of access: many people from rural areas or even third-world countries may not have the necessary infrastructure to use such software. While with the traditional procedures of mediation, there have been many camps in rural areas to introduce the people to the process and even train them to perform the same later. There would be a requirement of significant investment on the part of parties first to set up such infrastructure before they may gain access to such programs.[xvii]
- Lack of flexibility: when there is a human touch to the process of mediation, specific considerations and creative solutions may be achieved, which a computer may not be able to provide. A mediator may be able to provide a solution that may not have been part of the issues to help the parties settle the dispute, which allows the parties to satisfy all their interests. A machine would only be limited to the information it has been provided and thus would lack the skill to introduce a creative solution.20
The development of AI has brought about a lot of benefits to the process of mediation. The implementation of the same and for the people to trust it would take time. But soon, most of the procedures shall be automated and require minimal human interaction. More research needs to be done in this area to ensure that the programs being developed do not encounter any significant errors and bring about a smooth running method to settle disputes. But, one must also consider the fact that dispute resolution is a complex procedure that involves various factors like law, psychology, and society. The development of AI may not be able to take into account all these factors at once. The idea of AI being used to answer and ask a question may not be new to us. Still, the relationship-building part of the traditional mediation process may be a lot harder for it to achieve. AI may be the next big thing in the field of dispute resolution. Still, at present, it is hard to predict all the benefits and difficulties we may face with the integration of AI in mediation.
The need for AI or any form of online mediation has never been felt more than now. Due to the novel Corona Virus, many dispute resolution mechanisms came to a standstill which affected the parties greatly. AI would not have been affected by any such problem. As in this case, the parties would not have to meet and thus would still benefit from the dispute resolution mechanism remotely. Such programs are the need of the hour and would greatly benefit all persons in a dispute.
[i] Frankenfield Jake, Artificial Intelligence (AI) (Nov 5 2020, 4:14 pm), https://www.investopedia.com/terms/a/artificial-intelligence-ai.asp.
[ii] Emerj, What is Machine Learning (Nov 5 2020, 4:18 pm), https://emerj.com/ai-glossary-terms/what-is-machinelearning/.
[iii] JAMS Mediation Services, Mediation Defined: What is Mediation (Nov 5 2020, 4:20 pm),https://www.jamsadr.com/mediation-defined/.
[iv] Lodder R. Arno and Thiessen M. Ernest, Role of Artificial Intelligence in Online Dispute Resolution, p. 5-6 (Nov 5 2020, 5:20 pm), http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.97.9137&rep=rep1&type=pdf.
[v] Id. at 5-6.
[vi] Id. at 5-6.
[vii] Id. at 6-7.
[viii] Id. at 6-7.
[ix] Id. at 13.
[x] Id. at 13.
[xi] Id. at 13-14.
12Id. at 13-14.
[xii] Jain Sunaina, Pros and Cons of Artificial Intelligence in ADR, p. 2 (Nov 5 2020, 6:58 pm), found at https://viamediationcentre.org/readnews/NTE5/PROS-AND-CONS-OF-ARTIFICIAL-INTELLIGENCE-IN-ADR, last visited on 5-11-2020 at 6:58 pm.
[xiii] Id. at 2. 15Id. at 2.
[xiv] Id. at 2.
[xv] Id. at 2.
[xvi] Id. at 3.
[xvii] Id. at 3.
20Id. at 3.
*Maya Venkiteswaran is a student of Law in her 3rd year, at The National University of Juridical Sciences (NUJS), Kolkata. This blog entry won the first place in the Blog Writing Competition which was jointly organised by Ex Curia International and ODR Expo Tech Latin America.