Moderated by Ayelet Sela
“Online Dispute Resolution” has long been used as an umbrella term for a wide array of online procedures and technological tools that disputants and neutrals use to communicate and resolve disputes. As the field matures, it is now possible to envision more detailed taxonomies of ODR systems and discuss specific issues arising out certain classes of ODR. This forum is dedicated to discussing Technology Facilitated Resolution (TFR).
TFR refers to any ODR process in which technology plays the sole facilitative role. In other words, TFR processes do not involve a human third party; rather, technology plays the role of a third party. In Katsh & Rifkin’s terms, the fourth party plays the role of the third party. TFR processes include forms of automated resolution, such as problem diagnosis, algorithmic resolution, blind bidding, and if perceived broadly, even negotiation processes in which the platform offers a facilitative contribution to the process. A paradigmatic example of a TFR system is eBay’s Resolution Center, which processes 80% of its 60 million annual cases without any involvement of a human third party. Other examples include CyberSettle and Fair Outcomes.
TFR systems have many advantages. By making the dispute resolution process more structured, easier, faster and less expensive to complete, TFR can help improve redress and access to justice. Once a TFR system is built, it can be easily replicated and scaled, it functions consistently, it is unsusceptible to human error, and it can help overcome disputant biases and assist them in reaching Pareto-optimal resolutions. Importantly, in many classes of disputes, especially when the value of the dispute is too low to warrant paying a third party, TFR tools may be the only viable opportunity for redress. Because TFR systems require parties to describe disputes in standardized terms they also allow easier identification of dispute patterns, thereby facilitating both collective actions and policy and process improvements within institutions.
Alongside their many advantages, TFR systems also entail some risks. Automated systems may be seen as formal, mechanical, or depersonalized, and there may be fundamental differences in the way people react to them compared to humans. Some argue that certain types of TFR systems are prone to gaming. As in any other dispute resolution system, one needs to ensure that no bias is introduced into the design of the TFR system, and that there is sufficient transparency about the values enforced by the code.
Any discussion on TFR raises interconnected practical and philosophical questions. Practically, one may question the ability of software to perform dispute resolution tasks effectively. Normatively, one may question the moral desirability of using “machines” for resolving disputes. The two questions reflect the need to develop industry best practices, and possibly even an appropriate regulatory framework for TFR-driven ODR systems.
A recent article by Orna Rabinovich-Einy & Ethan Katsh,points to some of the basic tensions that the use of technology in ODR introduces, and thus serves as a good segue to this discussion:
new technologies represent more than a change in arena for the performance of dispute resolution processes; digital technology is transforming the very nature of these processes and changing their characteristics in ways that are bound to have an important impact on stakeholders and the organization alike… technology is by no means neutral and a particular software design reflects a preference for certain values over others… [However, t]he same software that promotes a particular value choice also makes design choices more visible, minimizes third party discretion due to enhanced structure, and allows for more ex-post study of the impact of design choices and quality control of decision-making where discretion is employed through data documentation and analysis.
…[B]ecause ODR processes are based on textual communication… [there is] an important benefit in terms of quality control over the process, its fairness and effectiveness… improper conduct, poor performance and problematic process design can be quite easily uncovered.
 Orna Rabinovich-Einy & Ethan Katsh, Lessons from Online Dispute Resolution for Dispute System Design, in: Online Dispute Resolution: Theory and Practice – A treatise on Technology and Dispute Resolution, Mohamed S. Abdel Wahab, Ethan Katsh & Daniel Rainey (eds.) 39, 50, 52 (2012). See also: Orna Rabinovich-Einy, Technology's Impact: The Quest for a New Paradigm for Accountability in Mediation, 11 Harv. Negot. L. Rev. 253, 274-276 (2006)
Ayelet is the director of Technology Facilitated Resolution at Modria Inc. Modria’s online platforms improve existing dispute resolution processes and access to justice in a variety of settings, including B2B, B2C, C2C, and governmental and court processes. Ayelet’s work focuses on designing automated resolution solutions and court-connected ODR systems. Ayelet is also a doctoral researcher at Stanford Law School, where she conducts an empirical evaluation of procedural justice in ODR systems in e-commerce and judicial settings. Her Master thesis, also gained from Stanford, explores barriers to adoption of court-connected mediation. Prior to joining Modria and Stanford, Ayelet was admitted as a lawyer to the Israeli Bar and clerked in the Israeli Supreme Court.
I'd like to welcome you again to the Technology Facilitated Resolution (TFR) forum! I'm excited to engage in a discussion with you on some of the opportunities and concerns that TFR brings about. One question that I've heard many ADR professionals raise, especially mediators, is whether TFR can foster party self-determination and empowerment. I personally believe that while TFR is not suitable for all kinds of disputes, for those disputes that can be "automatically" resolved, TFR can foster high levels of self-determination and empowerment. In fact, a structured, asynchronous, self-paced processes for collecting information about the dispute, educating the parties about their resolution options, and mutually exchanging information and offers may contribute to leveling the playing field between more/less powerful parties, allow more people the opportunity to "tell their story," and provide even greater certainty and a sense of process control.
I'd love to hear what you think!
I was very happy to discover on other CyberWeek forum discussions interesting ideas that apply to this forum. One great example is Bill Warters' great introductory learning module on communication theory and ODR, which touches upon implications of CMC on ADR tasks, differentiating by way of comparison between distinct forms of ODR. The section on automated resolution tools points out additional implication which I didn't mention in my earlier posts. Among TFR's benefits, it lists: privacy, quick and accessible, face saving, neutralises personal dynamics. Among TFR's risks it mentions: impersonal, lacks personal engagement of an ADR practitioner, current programs may be rigid, lacking intuition.
I'm not sure whether this list refers to more than blind bidding (which was mentioned as the single example of an automated resolution tool), but I definitely agree with the quote from Jelle van Veenen's work, cited towards the end of the module:
"...we have seen that online communication may miss features that are
available in face-to-face communication, but at the same time
gains others. For an online dispute resolution system to be effective,
it should not just copy an existing offline procedure. The missing
features will make it difficult to obtain good results, and opportunities
provided by the unique features of online communication are missed.
Rather, specific dispute resolution processes should be designed for
online dispute resolution."
I believe that this is exactly the promise of TFR.
Hi Ayelet - great to see you here!
I think two notions that might be helpful in considering where TFR might be helpful are what I call 'sorting' and 'staging'. The first invites people to consider how technology might be used to sort or sift disputes into different resolution categories. For example, eBay's system might quickly be able to identify when there is a delivery issue or a credit-card number mistake issue, a no-delivery, a "item-not-as-advertised' claim and so on - sifting those into separate 'boxes' in each of which a certain type of automated process takes over. It also might be designed to sift out, for example, disputes which seem to involve multiple attempts by buyer and seller to engage in communication (processes which might end up in a box for human involvement).
By 'staging', I mean that a TFR program might not necessarily have to offer a complete resolution system, but rather be placed to manage a certain stage of a larger system: The intake, the classification, and the post-op evaluation are just 'easy' stages that come to mind.
You might have better terms for these, of course (I'd be surprised if you didn't :-) ). In addition - you might envision TFR as intending only start-to-finish, 4th-as-3rd -party interventions.
Interested in hearing your thoughts!
Thank you for your insightful comments! Building on your suggested notions of "sorting" and "staging", I'd argue that TFR can offer even more dispute resolution functionality:
1. Case Profiling (a.k.a "problem diagnosis" or "sorting") - understanding the nature of the dispute in pre-defined discrete terms, which allow its subsequent routing to the appropriate resolution method / problem solver.
2. Resolution Generation:If the dispute space is well defined than a tightly defined case profiling process could enable automatic generation of resolution options, based on policy or law. These resolutions could be either directly enforced or serve as suggestions which a party may choose to adopt as offers to suggest to the other party.
3. Re-framing the Dispute and Structuring its Resolution: the language and structure of the case profiling process and resolution generation process can assist in re-framing the dispute in collaborative / neutral / interest-based terms, as well as anchor negotiation position on external standards, thereby contributing to the willingness of parties to resolve it. (This particular element requires high levels of diligence, neutrality and responsibility on the part of the system designer).
4. Case Routing (which you referred to as "staging", I believe): Assuming that a case could not be resolved in TFR modules (including structured, as opposed to free-text offer exchange), based on the case profiling it could be transferred to the next appropriate "human-facilitated" resolution method (free-form negotiation, mediation, arbitration, jury-process, judge etc).
5. Collective Action: Because case profiling and resolution behavior are recorded in the system, TFR opens new avenues for identifying a common problem that applies to a "class" and generate a resolution that could apply to all cases (thereby shortening the overall dispute resolution process time, and the resources required for it). If the institution providing the TFR process wishes too, TFR can also enable contacting members of the "class" who have not started a case, to remedy the problem for all members of the "class".
There are quite a few other TFR functions I can think of, but I believe that this initial list gives an idea of what a huge potential TFR has in improving access to redress and justice.
I think TFR is where ODR gets truly revolutionary. Ayelet is the leading thinker on the perimeters of TFR -- which really, is the same concept as the Fourth Party, as Ethan and Janet originally described it. Just with much more detail about the different techniques within it.
I can see putting the different components Ayelet describes can be organized into phases -- 1&3 are kind of intake/diagnosis, 4 is kind of triage/sorting/process design, and 2&5 focus on outcomes/solutions.
Most F2F mediators and arbitrators can't really wrap their heads around TFR. But I'd argue it's the part of ODR that will end up being the most consequential in changing the way people resolve their disputes. One day, TFR will be as big a part of dispute resolution as mediation and arbitration are today. On a volume basis, I bet TFR will handle 100x the volumes handled in F2F ADR in just a few years.
@Noam - absolutely - there is an incredible potential for both academic research and institutional improvement based on TFR data. Some large N research has been done in the past using eBay Resolution Center data, and hopefully, once TFR systems will be implemented in other organizations, academics will have much more to study!
@Colin - Thanks for your comment, I agree - in terms of volume, TFR is definitely expected to lead ODR usage statistics.
Colin's comment about traditional F2F mediators and arbitrators prompted me to point out that in addition to serving as a preliminary screening, diagnosis and routing stage to subsequent mediation and arbitration processes, TFR applications can also be incorporated into traditional (online or offline) dispute resolution processes. For example, mediators could incorporate a short questionnaire (similar to intake) as part of the mediation process, to diagnose in a structured way party differences in positions and interests. It reminded me of research that was conducted by Katalien Bollen, looking at the effect of incorporating an online intake stage into a F2F hierarchi....The research showed that adding an online intake stage to an F2F mediation had an equalizing effect on the parties' process experiences: while in F2F processes superiors had more favorable process experiences compared to subordinates, in hybrid processes subordinates reported equal levels of satisfaction, trust, uncertainty and procedural justice. Thus, I'd argue that TFR has the potential to contribute to- and strengthen, a broad range of existing dispute resolution processes.
@Colin - assuming that most of what the system at eBay falls under TFR, I think you've singlehandedly (well, almost :-) ) alread beaten that X100. 60 million disputes a year, if I remember correctly? Of course, they might not all qualify as 'disputes' in the eyes of a traditional mediator, but TFR is widening the scale of 'treat-as-dispute'.
The next frontier, I think, is X1000 . But you and Ayelet are probably way ahead of me on this...