Smartsettle’s algorithms enable their resolution assistance products (ONE and Infinity) to apply a new slant on Med-Arb. Rather than the sequential journey from mediation to failed mediation to arbitration, these systems (or rather mediator tools) offer a blending of mediation and arbitration.
e-Arbitration in ONE
ONE offers what is termed Visual Blind Bidding in which the parties, either with the assistance of a facilitator/mediator or by themselves, resolve their negotiation in a series of sessions in which they make both visible and secret bids to settle a numerical issue. Said numerical issue is represented by a blank in a Single Negotiating Framework, which is essentially the final agreement, except for the blank.
Offers are made by sliding flags along a horizontal scale. Agreement with ONE is reached in one of four ways:
- One party accepts the other party’s visible bid
- The secret bids overlap
- The Automatic Deal-Closer “arbitrates” a small gap
- The Expert Neutral Deal-Closer “arbitrates” a larger gap
Except when there is only one mutually agreed value, ONE uses a proprietary algorithm called “Reward Early Effort” (REE) to apportion the overlap or the remaining gap. REE is designed to promote collaboration and discourage parties from holding back in the bidding process. This process can be thought of as a type of AI-driven arbitration, which takes place to determine, through the binding conditions under which use of the system has been agreed, at which number within the range of either the overlap or the remaining gap is the settlement figure to be set. It might be thought that an obvious place to resolve the dispute would be at the midpoint between the last offers from each side. However, one problem with that method is that the last offer of the other party could be calculated by simply doubling the gap between that person’s last offer and the declared settlement. Even more serious, is that employing a simple split the difference formula tends to reward a non-collaborative party who holds back.
To overcome these problems, ONE applies the REE algorithm, which does precisely what it says and sets the final figure at a number within the range that is more favourable to the party who made more effort to settle, yet still within the agreed settlement range of both parties.
In the event that both parties fail to either overlap or come sufficiently close to enable the Automatic Deal-Closer to determine the final figure, then, should both parties agree, another algorithm, called the Expert Neutral Deal-Closer (END), comes into play to declare a final binding figure. When the END is used, three humans from an expert neutral roster are asked to declare a figure that they feel would be appropriate and fair to resolve the matter. The Expert Neutrals base their opinion on information within the Single Negotiating Framework, including any documents that may have been uploaded to support it.
In detail the Expert Neutrals are shown a range of values that covers the two final secret bids and extends an equal distance beyond each of the last secret bids to include the nearest visible proposal, but not extend beyond that figure. Imagine a history of bidding where the bids ranged from $47,000 to $62,000 with the final secret bids being $51,000 and $57,000. This is illustrated by the number line on the following illustration where one unit is equal to $1,000. The blue circles cover open bids and the yellow ones the secret bids.
In this case, the range of values shown to the Neutrals is as 47 to 61. 62 is excluded since it is 5 units away from the last secret bid of Party B, whereas 47 is only 4 units away from the last bid of 51 from Party A.
The Neutrals are not aware of the secret bids or visible proposals. All they know is that the range includes at least one of the visible proposals and that the secret bid gap is centred on the total range. The Neutral is instructed to choose one of these values that is the closest to fair in their expert opinion. With the participation of three neutrals, “fair” is deemed to be the middle value of the three.
If the middle value chosen by the Neutrals falls exactly between the accepted values or falls outside the accepted values, then that value becomes the outcome. Otherwise, Smartsettle uses the Reward Early Effort algorithm to favour the party who is closest to fair. Rounding if necessary is in favor of the party closest to fair.
In this case, suppose that the Neutral Panel identifies 55 as Fair.
In this example, Fair (55) falls between the accepted values. Since 55 is closer to 57 than 51 the final binding award is declared by Smartsettle to be 56.
Smartsettle ONE made history when, for the first time, AI was used for dispute resolution in a public court in England and Wales. Despite a contractual dispute running through the Small Claims Court for three months and the parties using the court’s own mediation service, no outcome had been reached. However once the parties were invited to use Smartsettle ONE the case settled in one hour. See here.
e-Arbitration in Infinity
“Infinity” is more appropriate in higher value disputes in which there are multiple issues and/or multiple parties. In these situations, it is more practical to use a system which represents all the issues as a complete package. A mediator using Infinity does not deny the parties the benefit of their usual skill and technique but rather, augments it with the benefit of a virtual assistant that can identify and offer solutions for the mediator to suggest to the parties more quickly and more accurately in relation to their interests than might otherwise be the case. Infinity is not about leaving the work to the machine but is very much dependent on the skills of a dispute resolution professional. The key is how the mediator, in discussion with each party, informs the system of the level of interest that each party has in, or importance they attribute to, each of the elements in a settlement. They also inform the system of the comparative importance each party places on one issue compared to each of the others.
One common application of Infinity is to resolve procurement disputes. Imagine a B2B supply chain between a supplier and a business customer. These parties have happily done business together for many years, but the customer is unhappy at the quality of the items sums in the latest delivery. The issues might be set as payment of some or all of monies due, return of some or all of the goods, replacement of some or all of the goods, commitment of the customer to further orders, discount of further orders and, perhaps the withdrawal of an adverse social media comment about the supplier. Once this is done, the system will evaluate by expressing a rating attributed to each of the various packages the mediator may post to the system. At any time, the mediator can ask Infinity to suggest other packages that it may feel, from the knowledge fed into it as to the relative and comparative interest levels of each party in each element of a package, are likely to be acceptable to both parties.
The parties present packages on a private basis to begin with and then, under the guidance of the mediator, can choose to share one or more of those packages with the other party. Both parties have the choice to accept one or more packages. Smartsettle checks for agreement at the end of each session. If multiple packages have been mutually accepted then Smartsettle uses the Reward Early Effort algorithm to determine which package becomes an agreement, which is a type of arbitration if you will. This agreement becomes a binding Baseline that determines how the parties will share the benefits of their negotiation. From that point, Infinity applies another algorithm called Maximize the Minimum Gain with which it looks for any value left on the table in order to offer an improved settlement that increases the benefits and value for both parties.
Both examples in this paper were of a machine arbitrated outcome but it's important not to lose sight of the fact that at all times the operation is under the control of the neutral and the parties they are assisting. The neutral may invite the parties to actively interact with the platform or simply use it privately to aid them in modelling and brainstorming suitable outcomes. If acting as a mediator they then can offer up one or more of these outcomes when brainstorming solutions with the parties, or, if acting as arbitrator, to guide him into an outcome to impose.
Graham Ross, Ernest Thiessen, November 7, 2020
Graham Ross is a UK lawyer and accredited mediator with over 30 years experience in IT and the law. Graham is the leading UK expert in the field of ODR (Online Dispute Resolution). He co-founded the UK's first ODR service, a blind bidding service called We Can Settle in 2000, and two years later the online mediation service The Mediation Room.
He is a Fellow of the National Center for Technology and Dispute Resolution at UMass (Amhurst)?.
Graham is a member of two Working Groups of the UK Civil Justice Council? ? being those on Alternative Dispute Resolution and on Online Dispute Resolution. Graham co-wrote the Report of the ODR Working Group which triggered the current moves? in the UK? to an online court.? Graham developed and runs the leading training course in applying technology to ADR ?at www.ODRTraining.com? ?
Ernest Thiessen is President & CEO of ICAN Systems Inc. He is responsible for the direction and management of the company as well as its government, industry and community affairs. Dr. Thiessen has led iCan’s research and development efforts for over two decades to implement the ICANS patent, creating the world’s first secure multiparty negotiation support system on the Internet. Thiessen has been a guest speaker at most of the Annual International Forums on Online Dispute Resolution since they began in 2002. Prior to founding iCan Systems Inc. in 1993, Thiessen worked for seventeen years as a consulting engineer and researcher in Canada and in Nepal. Thiessen received his PhD degree from Cornell University in 1993, majoring in Water Resource Systems Planning and Analysis in the School of Civil and Environmental Engineering. While at Cornell University, Thiessen invented an efficient methodology to provide comprehensive support to very complex negotiation problems. This methodology has been patented and is now embodied in both Smartsettle One (for the simplest possible cases) and Smartsettle Infinity (for the most difficult multivariate multilateral cases). Ernie is passionate about using technology for peacemaking and sustainable development.