Popular in Economics

Economics

Gender and Economics

University of California Santa Barbara Shelly Lundberg

Banks and Policy Organisations

Greening the financial sector

European Investment Bank Matteo Ferrazzi

Economic History

Gravity with Gravitas: A Solution to the Border Puzzle

Boston College James E. Anderson

Economic Inequalities

Visions of Inequality

City University of New York Branko Milanovic

Economics

School Finance Equalization Increases Intergenerational Mobility

Yale University Barbara Biasi

Economics

Olivier Blanchard on U.S. Pandemic-Era Inflation

Peterson Institute for International Economics Olivier Blanchard

Strategy-Proof Multi-Issue Mediation: An Application to Online Dispute Resolution

View more insights in Social Sciences - Economics - Behavioral economics

View more figures from University of Sydney

View the page of Onur Kesten

Read the Study

· Tools Cite Embed

Mediation, a key method of alternative dispute resolution, has become a multibillion-dollar industry globally. Online dispute resolution (ODR) providers utilize automated e-negotiation systems. Onur Kesten discusses a market design approach to create a framework for developing mediation mechanisms that are strategy-proof, efficient, and individually rational.

Image courtesy of interviewee. September 20, 2024

Log-in or Sign-up to Faculti
Currently viewing this subject insight as a guest. You have insight(s) remaining for this month.

Request more information on how to sign up as an Institution here

AI is building Key Quotes, Points, Transcriptions...

Export as PDF

This video discusses the concept of mediation, focusing on online dispute resolution (ODR). Mediation involves a neutral mediator helping parties resolve conflicts outside of traditional court litigation, which can be time-consuming and costly. ODR uses technology to facilitate mediation online, allowing parties to communicate through digital platforms. The prevalence of ODR is growing, especially in e-commerce, and some estimates suggest that major platforms resolve over a billion disputes annually.

However, existing ODR systems face challenges, including the potential for strategic manipulation where parties may distort information to achieve favorable outcomes. The author emphasizes the importance of designing ODR platforms based on economic principles to ensure fairness, efficiency, and protection against such gaming.

The ideal ODR system should be strategy-proof (encouraging honest reporting), efficient (providing optimal outcomes for both parties), and individually rational (respecting the parties' bargaining ranges). This video suggests that future research should focus on creating robust mechanisms that can handle asymmetric information and ensure that all participants are incentivized to act honestly. In conclusion, the text calls for the integration of economic design principles in the development of ODR platforms to improve user satisfaction and dispute resolution efficiency, ultimately aiming for better outcomes in conflict resolution.

Language:

Export as PDF

Export as PDF
Here are five key points from the video about mediation and online dispute resolution (ODR):

1. **Definition and Importance of Mediation**: Mediation is characterized as a dispute resolution process where an independent mediator assists parties in identifying options and negotiating agreements. It serves as an alternative to traditional litigation, which can be lengthy, expensive, and formal, often resulting in backlogged judicial systems.

2. **Advantages of Online Dispute Resolution**: ODR leverages technology to facilitate mediation, enabling parties to resolve disputes through online platforms. This shift has led to cost savings, increased efficiency, and accessibility, particularly for e-commerce disputes, with reports of significant financial savings for municipalities adopting ODR systems.

3. **Challenges in Current ODR Systems**: Existing ODR platforms face issues related to strategic manipulation, where parties may distort information to gain better outcomes. Many ODR systems do not account for economic principles that dictate human behavior, leading to potential systemic inefficiencies and unfair outcomes.

4. **Proposed Solutions Through Economic Design**: The video discusses using market design principles to create ODR protocols that are strategy-proof, efficient, and individually rational. These mechanisms would require participants to truthfully reveal their bargaining ranges, thereby reducing the possibility of gaming the system.

5. **Future Directions for ODR Development**: The ongoing evolution of ODR systems necessitates further research into robust mechanisms that accommodate asymmetric information and ensure ease of use for all participants. The focus should be on creating intuitive and user-friendly platforms that do not compromise essential complexities, aimed at improving fairness, transparency, and participant satisfaction in dispute resolution.

Export as PDF
Here are the key quotes from the video on mediation and online dispute resolution:

1. **Introduction to Mediation**:
- "This work is about mediation and in particular, online dispute resolution."

2. **Definition of Mediation**:
- "Mediation is a dispute resolution process in which an independent third party known as the mediator assists the parties to identify and assess options and negotiate an agreement to resolve their dispute."

3. **Contrast with Litigation**:
- "Traditionally, if you have a conflict, this conflict need to be taken up in a court... resolved through litigation... make a binding decision. This can be a very lengthy, expensive, frustrating and formal process."

4. **Efficiency of Mediation**:
- "A case that would take months... to resolve in a court can be solved through a few affordable mediation sessions."

5. **Rise of Online Dispute Resolution (ODR)**:
- "With the advancement of technology, mediation has moved to online platforms, giving rise to online dispute resolution systems, ODR systems."

6. **Concerns with Current ODR Systems**:
- "The problem is that they are lacking fundamental economic principles... they do not take parties strategic considerations into account."

7. **Importance of Incentives**:
- "One of the fundamental principles that we teach... is the fact that people respond to incentives."

8. **Strategic Manipulation**:
- "A negotiator has all the incentive in the world to distort the information that they're providing to the system to obtain a better outcome for themselves."

9. **Principles of Market Design**:
- "Market design... focuses on creating mechanisms or institutions that lead to desired outcomes in a marketplace, especially when private information or strategic behavior is involved."

10. **Key Properties of an Ideal ODR Platform**:
- "An ODR platform is ideal if it possesses three key properties: strategy proofness, inevitable efficiency, and individual rationality."

11. **Challenge of Non-overlapping Bargaining Ranges**:
- "A zone of mutual gain need not always exist, which means that you cannot solve every dispute."

12. **Central Role of Fairness**:
- "If you insist on fairness, then you should always use the central element... which doesn't favor each side."

13. **Role of Robust Algorithms**:
- "To do this, you need robust algorithms to ensure that strategic manipulation... doesn't undermine the intended fairness and efficiency of the resolution outcome."

14. **Implications for Legal Systems**:
- "The rise of online dispute resolution has major implications for legal systems worldwide, particularly in terms of reducing court backlogs, increased cost, saving time-wise, financially and efficiency."

These quotes encapsulate the main themes and important concepts presented in the video regarding mediation, online dispute resolution, and the economic principles that can improve these processes.

Export as PDF
Here are some articles that are similar in topic to the video, which focuses on mediation, online dispute resolution (ODR), and related economic considerations:

1. **"The Future of Online Dispute Resolution: An Overview"** - This article explores the evolution of ODR technologies and their potential impacts on the traditional legal system, including benefits and challenges.

2. **"Economic Models of Alternative Dispute Resolution"** - This piece delves into the economic principles behind mediation and arbitration processes, examining cost implications and efficiency considerations.

3. **"Mediation in the Digital Age: The Rise of Online Platforms"** - This article discusses the transition of traditional mediation practices to online platforms, detailing how technology is reshaping conflict resolution.

4. **"Strategies for Effective Online Dispute Resolution"** - This article outlines best practices for implementing ODR systems, focusing on the design of frameworks that encourage honest participation and mitigate manipulation.

5. **"Understanding Incentives in Dispute Resolution"** - A discussion on how incentive structures within ODR systems can influence negotiation outcomes, including insights from behavioral economics.

6. **"Challenges of Implementing Online Dispute Resolution Across Jurisdictions"** - This article examines the legal and logistical challenges faced by ODR platforms when operating in different legal environments.

7. **"User Experience in Online Dispute Resolution: Making Systems Intuitive"** - This piece emphasizes the importance of user-friendly design in ODR systems, particularly for those unfamiliar with legal processes.

8. **"Algorithmic Fairness in Online Dispute Resolution"** - A study investigating how algorithms can ensure fair outcomes in ODR, addressing concerns about biases and transparency in automated systems.

9. **"The Role of Technology in Modern Mediation"** - This article reviews various technological advancements that are being integrated into mediation processes, including AI-based tools for negotiation facilitation.

10. **"Toward a Better Future: Designing Effective Online Dispute Resolution Systems"** - This article outlines a framework for developing ODR systems with an emphasis on fairness, efficiency, and user engagement as key components.

These articles collectively highlight the intersection of technology, law, economics, and user experience in the convideo of mediation and online dispute resolution.

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:00.000 --> 00:01.290
This work is about mediation

00:01.290 --> 00:03.605
and in particular, online dispute resolution.

00:04.105 --> 00:06.285
To give you sort of some background on this topic,

00:06.475 --> 00:10.205
mediation is a dispute resolution process in which an

00:10.235 --> 00:13.125
independent third party known as the mediator,

00:13.505 --> 00:16.205
assist the parties to identify and assess options

00:16.345 --> 00:19.765
and negotiate an agreement to resolve their dispute.

00:20.225 --> 00:23.845
So traditionally, if you have a conflict, this conflict,

00:23.875 --> 00:25.565
need to be taken up, in a court.

00:25.585 --> 00:29.245
So it's resolved through litigation in which a court of law,

00:29.345 --> 00:30.805
is going to listen to the parties,

00:31.885 --> 00:33.525
evaluate their evidence, and ultimately

00:33.525 --> 00:34.645
make a binding decision.

00:34.915 --> 00:37.845
This can be a very lengthy, expensive,

00:38.325 --> 00:39.525
frustrating and formal process.

00:40.065 --> 00:42.565
And this is also evidenced in, by the fact

00:42.565 --> 00:44.685
that judicial systems of many countries

00:44.685 --> 00:47.925
around the world have a large backlog of cases, and,

00:47.925 --> 00:50.245
and efficiency concerns often motivate them

00:50.545 --> 00:53.605
to implement reforms, that focus on increased usage

00:53.605 --> 00:57.005
of alternative dispute resolution, processes

00:57.385 --> 01:00.165
and mediation, is often the preferred form

01:00.165 --> 01:02.485
of alternative dispute resolution,

01:02.485 --> 01:05.165
because of its, cost effectiveness time-wise,

01:05.595 --> 01:09.405
financially, its flexibility, its informality

01:09.745 --> 01:10.805
and confidentiality.

01:11.185 --> 01:12.965
So a case that would take months

01:13.185 --> 01:15.205
or several thousands of dollars

01:15.705 --> 01:17.965
to resolve in a court can be solved

01:17.965 --> 01:20.445
through a few affordable mediation sessions.

01:20.985 --> 01:23.805
And since the 1990s, a significant number

01:23.805 --> 01:26.925
of countries have implemented both, mandatory

01:26.945 --> 01:30.525
and voluntary mediation programs to improve the efficacy

01:30.825 --> 01:32.205
of their, legal systems.

01:32.465 --> 01:34.725
The study is online dispute resolution.

01:34.745 --> 01:37.765
So we're taking mediation to online,

01:37.985 --> 01:40.405
and I'm gonna refer to it as ODR.

01:40.585 --> 01:43.405
So, with the advancement of technology,

01:43.405 --> 01:46.725
mediation has moved to online platforms, giving rise

01:46.785 --> 01:50.325
to online, dispute resolution systems, ODR systems.

01:50.905 --> 01:54.405
So in an ODR system, parties interact

01:54.405 --> 01:56.045
through an online platform,

01:56.345 --> 02:00.005
and the mediator is usually, a patented software,

02:00.355 --> 02:04.525
also known as an, e negotiation system, ENS,

02:04.555 --> 02:08.525
that follows predetermined sets of rules, embodied,

02:08.525 --> 02:09.965
in a mechanized algorithm.

02:10.465 --> 02:14.685
So there are currently over 140 ODR platforms

02:15.115 --> 02:16.285
operating worldwide.

02:16.705 --> 02:18.805
Uh, and to mention, some,

02:18.825 --> 02:21.485
famous prominent ones are like smart saddle,

02:21.905 --> 02:23.645
square trade, cyber saddle.

02:23.815 --> 02:24.965
These are like the oldest

02:25.185 --> 02:27.245
and probably, the most prominent ones.

02:27.865 --> 02:30.685
Uh, and it is currently estimated that e-commerce

02:31.245 --> 02:34.525
platforms like eBay, PayPal, Uber, Amazon,

02:34.525 --> 02:38.645
they resolve more than a billion disputes, annually.

02:39.145 --> 02:41.045
Uh, CS governments are sort

02:41.045 --> 02:43.805
of increasingly adopting these systems, New York City,

02:44.025 --> 02:46.245
to give an example, reports

02:46.245 --> 02:48.325
that they have saved millions of dollars

02:48.815 --> 02:50.565
after switching to ODR

02:50.825 --> 02:52.965
for solving the disputes against the city.

02:53.425 --> 02:56.365
Uh, and many experts in the area believe that,

02:56.465 --> 02:58.725
online dispute resolution is

02:58.745 --> 03:00.405
The, inevitable future

03:00.505 --> 03:03.125
of dispute resolution in the new millennium.

03:03.385 --> 03:06.405
One key motivation for this is the existing system.

03:06.545 --> 03:10.005
So when we, analyze the existing ODR systems,

03:10.425 --> 03:12.285
the problem is that they are lacking

03:12.285 --> 03:13.845
fundamental economic principles.

03:14.145 --> 03:16.045
In particular, they do not take parties

03:16.725 --> 03:18.525
strategic considerations into account.

03:19.025 --> 03:20.285
Uh, the algorithms

03:20.285 --> 03:23.725
that they have engineered are designed based on the

03:23.725 --> 03:26.205
presumption that the disputes are going

03:26.205 --> 03:27.925
to truthfully reveal,

03:27.925 --> 03:31.005
their bargaining positions on whatever the dispute is.

03:31.425 --> 03:33.645
Uh, but as economists, you know, this is sort

03:33.645 --> 03:34.885
of unacceptable to us

03:34.885 --> 03:38.365
because one of the fundamental principles that we teach,

03:38.545 --> 03:39.725
we sort of something we,

03:39.785 --> 03:43.365
we first teach our students in our introductory classes is

03:43.365 --> 03:45.285
the fact that people respond to incentives.

03:46.025 --> 03:48.645
And if you fail to take that into account, you're going

03:48.645 --> 03:50.965
to often end up with unintended consequences.

03:51.705 --> 03:54.925
And in the context of, online dispute resolution,

03:55.445 --> 03:57.965
a negotiator has all the incentive in the world

03:58.025 --> 04:00.845
to distort the information that they're providing

04:00.845 --> 04:03.845
to the system to obtain a better, outcome

04:03.845 --> 04:05.645
to the resolution, for themselves.

04:06.145 --> 04:09.405
Uh, and, and this consideration is completely absent in the

04:09.405 --> 04:11.765
current systems, in the current ODR systems.

04:12.465 --> 04:16.045
Uh, a platform that would be prone to gaming,

04:16.745 --> 04:18.645
may produce systematically unfair

04:18.665 --> 04:22.725
and inefficient outcomes, defeating the very purpose,

04:22.905 --> 04:25.045
on which, these platforms are designed.

04:25.425 --> 04:28.005
Uh, and it is quite possible that a platform

04:28.005 --> 04:32.965
that is gameable, may declare an impasse, for an

04:32.965 --> 04:34.765
otherwise resolvable dispute.

04:35.365 --> 04:37.165
Existing research also supports this view.

04:37.225 --> 04:41.485
It shows that strategic gaming in online platforms can

04:41.545 --> 04:44.725
result in, more escalation than resolution.

04:45.185 --> 04:48.165
So all of these issues set the stage for the approach

04:48.165 --> 04:52.245
that we take, in our work, namely achieving immunity

04:52.385 --> 04:54.285
to gaming through strategic proof,

04:54.635 --> 04:58.725
through strategic proofness, by pausing the problem,

04:58.725 --> 05:02.245
through the lens of economic engineering, what we call as,

05:02.305 --> 05:04.925
market design or mechanism design.

05:05.265 --> 05:07.605
Uh, and, and to mention one more problem

05:07.715 --> 05:09.805
with the current platforms,

05:09.805 --> 05:11.325
which also motivates our study.

05:11.825 --> 05:15.525
Um, the algorithms that these systems are using, are,

05:15.545 --> 05:17.485
are sort of opaque algorithms.

05:17.635 --> 05:18.765
It's unknown to users.

05:19.345 --> 05:22.205
Uh, so some platforms like the cyber saddle that is used

05:22.205 --> 05:25.285
by New York City chooses not to disclose their algorithm

05:25.285 --> 05:28.205
because of their patent infringement,

05:28.205 --> 05:31.925
concerns other platforms, famous platforms such as,

05:31.935 --> 05:35.725
smart Settle, also adopt this, black box, approach

05:35.915 --> 05:38.885
because their algorithms involve, sophisticated

05:38.985 --> 05:42.205
and unintuitive, inte integer optimization techniques.

05:42.625 --> 05:44.845
So, designers of smart settle, for instance,

05:44.845 --> 05:48.085
they would openly state that their multi-issue,

05:48.405 --> 05:52.125
renegotiation system cannot be used to its full potential,

05:52.385 --> 05:54.885
by inexperienced users,

05:54.885 --> 05:56.965
without the assistance of a facilitator.

05:57.585 --> 06:00.325
So, when looked at from an economic design perspective,

06:00.615 --> 06:03.245
these issues, raise red flags, you know,

06:03.245 --> 06:04.765
raise serious concerns for us,

06:04.955 --> 06:07.965
because in our area, it is widely acknowledged

06:08.225 --> 06:11.645
and already, commonly accepted that, the,

06:11.645 --> 06:15.165
the simplicity and the transparency of the underlying

06:15.725 --> 06:18.565
mechanics of an algorithm are very crucial

06:18.825 --> 06:20.925
for its effectiveness, for its efficacy.

06:21.025 --> 06:24.365
In economics, for example, we have this rich literature,

06:24.365 --> 06:28.205
of mechanism and market design that can help, to deal

06:28.205 --> 06:29.205
with all these issues.

06:29.705 --> 06:33.525
Uh, market design in a nutshell, is a field of economics

06:33.525 --> 06:36.125
that focuses on creating mechanisms

06:36.145 --> 06:38.405
or institutions, that lead

06:38.405 --> 06:40.485
to desired outcomes in a marketplace,

06:40.575 --> 06:42.965
especially when private information

06:43.025 --> 06:45.005
or strategic behavior is involved.

06:45.945 --> 06:48.525
And this literature is, is in, is based on decades

06:48.665 --> 06:50.485
of study on human incentives,

06:50.865 --> 06:53.365
and provide us, with the foundations

06:53.745 --> 06:56.965
for designing sound institutions for practical use.

06:57.265 --> 07:00.325
So, in fact, these studies have led to several Nobel Prizes,

07:00.585 --> 07:03.885
in our field, based on several applications,

07:03.945 --> 07:06.685
for example, applications about auction design.

07:06.985 --> 07:09.765
How should you design, an auction to sell,

07:10.145 --> 07:11.325
radio spectrum?

07:11.545 --> 07:14.805
How should you design an auction to sell,

07:14.995 --> 07:16.645
sponsored keywords online?

07:16.745 --> 07:19.285
How do you use it to sell oil timber?

07:19.625 --> 07:22.645
So all of these are sort of subject of our field and,

07:22.645 --> 07:24.845
and sound mechanisms have been developed

07:24.845 --> 07:26.885
and have been adopted, in practice.

07:27.225 --> 07:29.845
How, other examples are like student assignment,

07:29.985 --> 07:32.205
how students should be assigned to public schools

07:32.345 --> 07:35.325
or colleges or, in, in medicine,

07:35.665 --> 07:38.045
how organ transplantation should take place.

07:38.185 --> 07:40.565
How should it be organized in the most efficient way

07:40.905 --> 07:42.685
to maximize, life save?

07:42.865 --> 07:46.485
So our study is inspired by this type of literature,

07:46.485 --> 07:48.885
which provides us with the key set of tools

07:49.345 --> 07:50.405
to tackle the design

07:50.405 --> 07:54.205
of an ideal online dispute resolution platform, and,

07:54.225 --> 07:57.365
and to sort of provide the objectives such as fairness,

07:57.935 --> 08:00.965
efficiency, and immunity, to gaming much

08:00.965 --> 08:03.525
of the traditional mechanism design, approach

08:03.525 --> 08:07.165
to bargaining, which is pioneered by a seminal work

08:07.165 --> 08:10.365
of our Nobel, prize winning economist, Roger Myers.

08:10.365 --> 08:14.245
And then his coauthor, mark, would actually equip,

08:14.385 --> 08:16.285
negotiators with quasi linear,

08:16.355 --> 08:18.245
what we call quasi linear utilities.

08:18.245 --> 08:21.605
So you care about, an an object that you receive,

08:21.605 --> 08:25.045
and also the monetary transfer, that you, that you get.

08:25.545 --> 08:27.845
Uh, so, and in disputes, often money,

08:27.845 --> 08:30.805
is an important issue that comes up that we see,

08:30.905 --> 08:32.565
but it's rarely the only issue.

08:32.625 --> 08:34.925
So for this reason, in our framework,

08:35.155 --> 08:38.245
what we do is we adopt a more general framework,

08:38.255 --> 08:41.285
based on what we call an oral preference,

08:41.315 --> 08:43.805
reporting approach, where, we're going

08:44.005 --> 08:46.005
to ask the negotiators,

08:46.065 --> 08:48.285
to reveal only their bargaining ranges.

08:48.465 --> 08:51.245
So, for example, tell me, the range

08:51.265 --> 08:53.285
of negotiable alternatives for you,

08:53.545 --> 08:56.525
rather than giving the sophisticated information about like

08:56.525 --> 08:57.805
your utility functions,

08:58.075 --> 08:59.965
your full fledged cardinal preferences

09:00.025 --> 09:02.725
and things, sort of, that is, quite abstract,

09:02.865 --> 09:03.925
for the layman.

09:04.345 --> 09:07.845
Um, and these are sort of the three advantages, one

09:07.845 --> 09:09.685
of the three advantages of, of our approach

09:09.835 --> 09:13.885
with this oral preference reporting language, we are able

09:13.885 --> 09:17.765
to, develop detail free, mechanisms, which is,

09:17.765 --> 09:21.325
again, hailed by Nobel Prize winning economists as a must,

09:21.745 --> 09:24.165
property that needs to be satisfied

09:24.465 --> 09:25.685
for practical success.

09:25.815 --> 09:29.325
These detail free mechanisms provide, robust incentives

09:29.625 --> 09:31.405
to participants, by,

09:31.505 --> 09:33.805
by imposing the fundamental property

09:34.105 --> 09:36.445
of strategy proofness, in our area.

09:36.625 --> 09:39.085
So an online dispute resolution,

09:39.085 --> 09:41.245
mechanism should be strategy proof.

09:41.665 --> 09:44.365
It should be in the best interest of your participants

09:44.665 --> 09:48.365
to reveal their positions, truthfully, in an honest way,

09:48.515 --> 09:51.405
without having any attempts to game the system.

09:51.945 --> 09:54.885
So this sort of sets up the backbone of our,

09:54.885 --> 09:59.125
formal model, where, we have two negotiators,

09:59.145 --> 10:01.245
who are in a dispute over two issues.

10:01.705 --> 10:04.085
Uh, for simplicity, I'm going to refer

10:04.085 --> 10:05.845
to these two issues as a main issue

10:06.105 --> 10:07.445
and a supplementary issue,

10:07.465 --> 10:10.925
but it can be generalized to any number of issues, as long

10:10.925 --> 10:12.285
as you have more than one issue.

10:12.525 --> 10:15.325
A crucial insight, in, in this work is

10:15.325 --> 10:19.725
that is the embodiment of a principle, that is,

10:19.725 --> 10:22.445
that has been sort of repeatedly brought up

10:22.465 --> 10:23.805
by negotiation experts.

10:24.145 --> 10:26.965
So negotiation experts, if you take like, one

10:26.965 --> 10:30.045
of the sort of bestseller books, they would say

10:30.045 --> 10:33.885
that success in mediation lies in the party's ability

10:34.305 --> 10:38.165
to expand the pie and find integrative win-win outcomes.

10:38.745 --> 10:41.525
Uh, so, so parties should be able to,

10:41.555 --> 10:42.925
make concessions to each other.

10:43.425 --> 10:46.725
And, and this necessitates the sort of re,

10:46.785 --> 10:50.845
the requirement for multi-issue negotiation and log rolling.

10:51.665 --> 10:54.685
Um, so these issues, in our, sort

10:54.685 --> 10:58.245
of formal model is represented by, discrete

10:58.245 --> 11:00.725
and bounded sets of resolution alternatives.

11:01.025 --> 11:04.125
And each alternative represents a settlement,

11:04.125 --> 11:05.525
for a particular issue.

11:06.465 --> 11:07.805
So we assume, for example,

11:07.835 --> 11:11.405
that each negotiator has a commonly known ranking

11:11.465 --> 11:12.565
of the alternatives,

11:12.745 --> 11:15.525
and these alternatives are diametrically opposed.

11:15.545 --> 11:18.285
So to simplify what I'm trying to say, you know,

11:18.285 --> 11:19.845
suppose we're trying to divide assets.

11:20.265 --> 11:23.325
Uh, each one of us prefers to get the larger share

11:23.445 --> 11:24.485
of the assets, okay?

11:24.485 --> 11:26.445
So our preferences are diametrically opposed.

11:26.665 --> 11:29.925
You can think of it as a zero sum game, in each issue.

11:30.465 --> 11:34.125
Um, but what is sort of at the crux of this matter is that

11:34.675 --> 11:37.325
each negotiator's bargaining range is

11:37.325 --> 11:38.445
their private information.

11:38.505 --> 11:41.405
You know, when I'm gonna walk away from negotiation, it's,

11:41.405 --> 11:42.925
it's what I know privately,

11:43.305 --> 11:46.485
and I don't have any interest in revealing that information

11:46.585 --> 11:48.085
to you or the mediator.

11:48.275 --> 11:53.125
Okay? So this information, is, is sort of in economics,

11:53.205 --> 11:55.445
what we refer as the type of the agent,

11:55.785 --> 11:58.205
and you cannot force them to reveal their type.

11:58.465 --> 12:02.165
You have to, design these mechanisms in a smart way,

12:02.435 --> 12:05.565
such that once they see the mechanism, they will realize

12:05.595 --> 12:09.125
that it's in their best interest to reveal, their types,

12:09.175 --> 12:12.045
truthfully, without attempting any manipulation.

12:12.095 --> 12:14.125
Other, observation in the context

12:14.305 --> 12:15.725
of dispute resolution is

12:15.725 --> 12:18.285
that depending on these type realizations,

12:18.745 --> 12:22.285
the bargaining ranges of your negotiators may not overlap.

12:22.705 --> 12:26.565
So a zone of mutual gain need not always exist, which,

12:26.565 --> 12:28.925
which means that you cannot solve every dispute.

12:29.065 --> 12:31.685
You know, there's no such thing as solving,

12:31.685 --> 12:33.445
every dispute successfully.

12:34.145 --> 12:38.685
Um, so our hope in this study, is to actually

12:39.445 --> 12:41.685
identify the ideal ODR platform.

12:42.105 --> 12:45.805
Uh, and, and in our sort of formal setting, we, we sort of

12:46.325 --> 12:47.885
abstractly say that,

12:48.185 --> 12:51.805
and ODR platform is ideal if it possesses

12:51.815 --> 12:53.005
three key properties.

12:53.705 --> 12:56.365
Uh, and these are strategy proofness.

12:56.705 --> 12:59.165
Uh, the, the incentive issue that we've been talking about,

12:59.225 --> 13:00.805
it has to be immune to gaming.

13:01.725 --> 13:04.285
Whatever information that you are requiring your

13:04.285 --> 13:07.645
participants to reveal, they should reveal it truthfully.

13:08.105 --> 13:10.485
And the second property, as economists for us,

13:10.545 --> 13:12.245
is inevitable efficiency.

13:13.165 --> 13:15.525
Whatever resolution your platform is offering,

13:15.985 --> 13:18.765
you should not be able to find an alternative resolution

13:18.915 --> 13:20.645
that is better for both parties.

13:20.995 --> 13:24.205
Okay? So you should pick one of the optimal, outcomes.

13:24.585 --> 13:25.645
And the last one,

13:25.835 --> 13:28.365
what we technically call individual rationality,

13:28.655 --> 13:32.325
which simply means that, the offers that you are making

13:32.345 --> 13:34.925
to these negotiators should never lie

13:35.155 --> 13:37.045
outside their bargaining ranges.

13:37.645 --> 13:39.085
Whatever information they tell you,

13:39.265 --> 13:41.805
you should respect their bargaining ranges, you should not,

13:42.145 --> 13:44.205
sort of jump, try to jump the gun.

13:44.825 --> 13:48.645
Um, so in, in the, in our study, what we first showed,

13:48.645 --> 13:52.765
that not all disputes are necessarily solvable by an ideal,

13:52.945 --> 13:55.005
by an ideal ODR protocol.

13:55.505 --> 13:58.485
So it all hinges on party's ability

13:58.505 --> 14:00.525
to make concessions on the issues.

14:00.865 --> 14:03.645
So in the legal terminology, for instance, this is

14:03.645 --> 14:06.565
what is referred as the quid pro quo property.

14:06.945 --> 14:10.405
So the parties should be able to trade favors, and,

14:10.425 --> 14:14.885
and if the negotiators underlying preferences satisfy this

14:15.005 --> 14:17.405
quid pro quo property, which we formally sort

14:17.405 --> 14:19.325
of capture in the paper, then we show

14:19.325 --> 14:22.525
that you can actually construct an ideal mechanism

14:22.665 --> 14:24.365
and algorithm that is going

14:24.365 --> 14:27.405
to successfully solve these disputes,

14:27.405 --> 14:29.525
meeting those three key properties.

14:30.105 --> 14:33.485
Um, so what we do in the paper is actually,

14:33.505 --> 14:36.885
we provide a full characterization of these,

14:36.945 --> 14:39.845
online dispute resolution protocols,

14:39.845 --> 14:41.165
that are immune to gaming.

14:41.605 --> 14:45.125
IE they are strategy proof that are efficient and,

14:45.225 --> 14:46.805
and they're individually rational,

14:46.995 --> 14:50.165
providing the negotiators,

14:50.255 --> 14:52.085
admissible, proposals.

14:52.425 --> 14:54.685
So these are, what we in the paper refer

14:54.685 --> 14:56.525
as these log rolling mechanisms.

14:56.865 --> 14:58.885
Uh, intuitively speaking,

14:58.885 --> 15:02.605
these mechanisms recommend bundles that involve compromises,

15:02.665 --> 15:04.205
in the issues in the dispute.

15:04.385 --> 15:07.405
So, for example, if the res, if the resolution package

15:07.435 --> 15:11.245
that you are offering to a negotiator, involves,

15:11.265 --> 15:14.445
an alternative that is favorable to one negotiator,

15:14.795 --> 15:18.045
then from the other issue, you have to pick an alternative

15:18.045 --> 15:20.725
that is favorable to the opposite negotiator.

15:20.985 --> 15:22.525
So after you have to only bring these

15:22.525 --> 15:23.845
concession options together.

15:24.465 --> 15:27.765
Uh, and, and sort of on the mathematical side, on the sort

15:27.965 --> 15:31.405
of formal side of things, we actually show that these type

15:31.405 --> 15:34.525
of bundles, these type of log roll, what we call

15:34.525 --> 15:35.725
as log rolling bundles,

15:35.835 --> 15:37.725
have a beautiful mathematical structure.

15:37.905 --> 15:41.085
So they satisfy this mathematical property,

15:41.085 --> 15:44.965
what is called the semi latice property, in set theory.

15:45.225 --> 15:48.325
So any sort of those two bundles, you take two of them,

15:48.585 --> 15:51.725
you can actually construct, an upper bound a bundle

15:51.795 --> 15:54.045
that is better than both of them, which means

15:54.045 --> 15:57.285
that you always have these external optimal members,

15:57.415 --> 16:01.005
which gives us this beautiful sort of, incentive result,

16:01.305 --> 16:02.685
in our characterization.

16:03.265 --> 16:05.605
Uh, and, and, and based on the structure,

16:05.705 --> 16:08.725
we actually propose a general class of,

16:09.045 --> 16:10.245
resolution algorithms.

16:10.665 --> 16:14.045
And, and we show that there are multiple ways one can

16:14.045 --> 16:17.445
implement these resolution protocols, which is the kind

16:17.445 --> 16:20.805
of flexibility we would like to give these online platforms.

16:21.105 --> 16:22.405
You can do it simultaneously.

16:22.865 --> 16:25.165
You can offer to make them simultaneous offers.

16:25.585 --> 16:26.925
You can do it sequentially.

16:27.025 --> 16:30.325
You can approach one negotiator, tell them, you know, the,

16:30.325 --> 16:31.725
sort of the available options,

16:31.785 --> 16:35.885
or ask them to sort of provide a menu of options

16:35.885 --> 16:37.245
that is acceptable to them.

16:37.545 --> 16:39.885
And then, taking these options, the,

16:39.945 --> 16:43.565
the mediator can add some options of their preference

16:43.785 --> 16:45.565
and then turn to the other negotiator

16:45.665 --> 16:47.565
and ask them, you know, would you,

16:47.565 --> 16:48.965
would you pick one of these options?

16:49.065 --> 16:51.765
Do you find any of these options, acceptable?

16:51.785 --> 16:54.325
So it can be simultaneously, it can be sequential,

16:54.665 --> 16:56.805
it can be implemented in a centralized fashion

16:56.945 --> 17:00.685
or a decentralized fashion, a synchronously online,

17:00.745 --> 17:03.725
you know, in case parties are not participating in the

17:03.725 --> 17:06.645
process at the same time, which is a sort of a comfort

17:06.835 --> 17:09.765
that the online platforms are providing their users.

17:10.425 --> 17:14.845
Um, you can do a shortlisting type of, resolution.

17:14.865 --> 17:18.205
For example, you can have one negotiator offer a shortlist

17:18.305 --> 17:20.205
of, resolution options

17:20.225 --> 17:22.405
as negotiable solutions for the dispute.

17:22.785 --> 17:25.525
And the mediator would then communicate these options

17:25.785 --> 17:28.885
to the other negotiator who then, would choose,

17:28.945 --> 17:31.165
her favorite, option from the list.

17:31.345 --> 17:34.445
So, so there is a lot of flexibility in these, class

17:34.445 --> 17:37.245
of mechanisms, while also ensuring

17:37.435 --> 17:41.565
that our full deserter of strategy, proofness efficiency

17:41.665 --> 17:44.285
and individual rationality, is met.

17:44.765 --> 17:46.685
Remember, we also assume that,

17:46.705 --> 17:48.725
the underlying preferences are compatible

17:48.755 --> 17:51.325
with the squid pro quo assumption that the,

17:51.345 --> 17:53.645
the negotiators are able to, make,

17:53.675 --> 17:55.285
make concessions among themselves.

17:55.545 --> 17:58.045
And, and this allows us to always,

17:58.045 --> 18:00.525
make a recommendation from the set of,

18:00.585 --> 18:01.645
log rolling bundles.

18:01.905 --> 18:04.805
The, the set, consists of optimal elements.

18:04.985 --> 18:08.285
So this is not something we would like to sort of,

18:08.425 --> 18:10.245
of course, implement in practice,

18:10.305 --> 18:13.245
but from the sort of mathematical beauty of the theory,

18:13.825 --> 18:16.925
you can actually, obtain if you, if you wanted

18:16.925 --> 18:20.285
to favor one side of the dispute, you can actually obtain,

18:20.465 --> 18:24.005
the sort of algorithms that would, favor one side

18:24.025 --> 18:27.765
of the dispute while picking, you know, those concessional,

18:28.105 --> 18:30.805
outcomes that is in the best interest

18:30.825 --> 18:32.645
of one side in a systematic way.

18:32.985 --> 18:34.685
Uh, but of course, we don't do that.

18:34.705 --> 18:36.325
We don't wanna do that as, and, and,

18:36.345 --> 18:39.925
and we focus in the study at the, central element

18:39.985 --> 18:43.605
of our class of mechanisms, which we refer as a,

18:43.925 --> 18:45.805
constrained, shortlisting, mechanism.

18:45.865 --> 18:48.645
And we show that this actually satisfies a very

18:48.645 --> 18:50.365
natural fairness requirement.

18:50.625 --> 18:51.805
If you insist on fairness,

18:52.115 --> 18:54.605
then you should always use the central element,

18:54.975 --> 18:58.165
which is sort of anonymous, which doesn't favor each size.

18:58.385 --> 18:59.845
It is unbiased and,

18:59.845 --> 19:03.405
and can be sort of very easily implemented in practice.

19:03.585 --> 19:05.645
So that sort of forms the fundamental,

19:06.025 --> 19:07.485
proposal of the paper.

19:07.925 --> 19:10.925
I would say that as online dispute resolution,

19:10.925 --> 19:14.285
systems become more prevalent around the world,

19:14.395 --> 19:17.805
mediation, the traditional mediation is being integrated

19:17.915 --> 19:20.125
into online platforms where disputes,

19:20.285 --> 19:23.085
particularly those arising from e-commerce

19:23.385 --> 19:25.885
or other digital tra transactions, as well

19:25.885 --> 19:28.685
as offline disputes, are resolved through,

19:28.745 --> 19:29.805
digital means.

19:30.225 --> 19:32.605
And the, the idea of automating,

19:32.605 --> 19:35.685
the negotiation process, simplifies decision making

19:35.745 --> 19:37.485
and speeds up the resolution process.

19:38.065 --> 19:42.205
So it it, but to do this, you need robust algorithms

19:42.385 --> 19:45.725
to ensure that, strategic manipulation, you know, the,

19:45.745 --> 19:48.765
the ga, the non-game ability of the system,

19:48.915 --> 19:51.485
doesn't undermine the intended fairness

19:51.625 --> 19:54.685
and efficiency, of the, of the resolution outcome.

19:54.965 --> 19:56.205
Remember the incentives property.

19:56.505 --> 20:00.245
If you don't have good incentives, then no matter you know,

20:00.265 --> 20:02.805
how good your algorithm is in terms of efficiency

20:02.805 --> 20:05.205
and fairness, people are going to distort it,

20:05.425 --> 20:08.005
and the outcome is not going to be fair or efficient.

20:08.265 --> 20:11.485
That's why we argue that a market design approach

20:11.575 --> 20:15.165
where the principles of economics has to be sort

20:15.165 --> 20:18.285
of at the heart of this engineering problem,

20:18.545 --> 20:21.205
is necessary to improve the efficiency

20:21.225 --> 20:24.365
and fairness of, existing online dispute,

20:24.605 --> 20:25.725
resolution, problems.

20:26.225 --> 20:28.245
Uh, and the, what is the implication of this?

20:28.275 --> 20:29.805
Well, the rise

20:29.865 --> 20:32.565
of online dispute resolution has major implications

20:32.585 --> 20:36.285
for legal systems, worldwide, in particularly in terms

20:36.285 --> 20:39.125
of reducing court backlogs, increased cost,

20:39.125 --> 20:42.085
saving time-wise, financially and efficiency.

20:42.265 --> 20:45.165
And governments and courts must consider how

20:45.165 --> 20:48.685
to integrate these systems, while ensuring transparency

20:48.865 --> 20:52.285
and fairness, as well as, good incentive properties.

20:52.585 --> 20:56.925
So understanding incentives is extremely important, for,

20:57.345 --> 21:01.285
having the ultimately desirable, outcomes,

21:01.285 --> 21:04.005
in online dispute resolution, processes,

21:04.705 --> 21:06.325
for, for future.

21:06.625 --> 21:09.405
So our, our studies is, I believe,

21:09.585 --> 21:11.845
in market design in our field is the first

21:11.865 --> 21:13.845
to tackle this problem through,

21:14.165 --> 21:15.765
economic design, perspectives.

21:15.985 --> 21:19.045
Uh, but for future, there's more research that is necessary

21:19.065 --> 21:22.405
for designing robust mechanisms that resist gaming,

21:22.565 --> 21:26.085
particularly in situations where the parties are going

21:26.085 --> 21:29.365
to have as symmetric information or as symmetric experience.

21:29.705 --> 21:32.085
So, future studies should focus on ensuring

21:32.085 --> 21:35.125
that all participants, regardless of their background

21:35.145 --> 21:38.965
or experience, are incentivized to participate truthfully,

21:39.385 --> 21:42.045
giving the information, that is,

21:42.045 --> 21:43.485
that is truthful and honest.

21:43.905 --> 21:47.045
So, given that many ODR systems, the currently

21:47.565 --> 21:50.445
existing systems are difficult for inexperienced people

21:50.505 --> 21:53.725
to navigate without additional help, research,

21:53.725 --> 21:56.645
should focus on making these systems more intuitive

21:56.745 --> 21:59.885
and user-friendly without compromising on the

21:59.885 --> 22:01.725
complexity, needed.

22:01.785 --> 22:04.125
For fairness to sum up,

22:04.275 --> 22:08.125
designing successful online dispute resolution platforms is

22:08.165 --> 22:09.285
a new research area.

22:09.865 --> 22:11.645
And I do sincerely hope

22:11.645 --> 22:15.565
that the industry practitioners would involve more economic

22:15.925 --> 22:19.205
designers to help, build platforms that are simple,

22:19.675 --> 22:21.965
more efficient, fair, transparent,

22:22.265 --> 22:24.445
and that give their participants the right

22:24.445 --> 22:25.645
incentives to act.

22:25.945 --> 22:28.725
Uh, truthfully, this should help increase the number

22:28.745 --> 22:31.205
of disputes successfully, result,

22:31.205 --> 22:33.045
and quickly result while,

22:33.045 --> 22:35.365
boosting the user satisfaction,

22:35.365 --> 22:38.605
and ultimately, making the world a happier place.

00:00  - This work is about mediation and in particular, online dispute resolution. To give you sort of some background on this topic, mediation is a dispute resolution process in which an independent third party known as the mediator, assist the parties to identify and assess options

00:16  - and negotiate an agreement to resolve their dispute. So traditionally, if you have a conflict, this conflict, uh, need to be taken up, uh, in a court. So it's resolved through litigation in which a court of law, uh, is going to listen to the parties, evaluate their evidence, and ultimately make a binding decision. This can be a very lengthy, expensive, uh, frustrating and formal process. And this is also evidenced in, uh, by the fact

00:42  - that judicial systems of many countries around the world have a large backlog of cases, and, and efficiency concerns often motivate them to implement reforms, uh, that focus on increased usage of alternative dispute resolution, uh, processes and mediation, uh, is often the preferred form of alternative dispute resolution, uh, because of its, uh, cost effectiveness time-wise, financially, uh, its flexibility, its informality and confidentiality.

01:11  - So a case that would take months or several thousands of dollars to resolve in a court can be solved through a few affordable mediation sessions. And since the 1990s, uh, a significant number of countries have implemented both, uh, mandatory and voluntary mediation programs to improve the efficacy of their, uh, legal systems. The study is online dispute resolution. So we're taking mediation to online, uh,

01:37  - and I'm gonna refer to it as ODR. So, with the advancement of technology, uh, mediation has moved to online platforms, uh, giving rise to online, uh, dispute resolution systems, ODR systems. So in an ODR system, uh, parties interact through an online platform, and the mediator is usually, uh, a patented software, also known as an, uh, e negotiation system, ENS, uh, that follows predetermined sets of rules, uh, embodied, uh, in a mechanized algorithm.

02:10  - So there are currently over 140 ODR platforms operating worldwide. Uh, and to mention, uh, some, uh, famous prominent ones are like smart saddle, square trade, cyber saddle. These are like the oldest and probably, uh, the most prominent ones. Uh, and it is currently estimated that e-commerce platforms like eBay, PayPal, Uber, Amazon, they resolve more than a billion disputes, uh, annually.

02:39  - Uh, CS governments are sort of increasingly adopting these systems, uh, New York City, to give an example, uh, reports that they have saved millions of dollars after switching to ODR for solving the disputes against the city. Uh, and many experts in the area believe that, uh, online dispute resolution is The, uh, inevitable future of dispute resolution in the new millennium.

03:03  - One key motivation for this is the existing system. So when we, uh, analyze the existing ODR systems, the problem is that they are lacking fundamental economic principles. In particular, they do not take parties strategic considerations into account. Uh, the algorithms that they have engineered are designed based on the presumption that the disputes are going to truthfully reveal, uh,

03:27  - their bargaining positions on whatever the dispute is. Uh, but as economists, you know, this is sort of unacceptable to us because one of the fundamental principles that we teach, we sort of something we, we first teach our students in our introductory classes is the fact that people respond to incentives. And if you fail to take that into account, you're going to often end up with unintended consequences. And in the context of, uh, online dispute resolution,

03:55  - a negotiator has all the incentive in the world to distort the information that they're providing to the system to obtain a better, uh, outcome to the resolution, uh, for themselves. Uh, and, and this consideration is completely absent in the current systems, in the current ODR systems. Uh, a platform that would be prone to gaming, uh, may produce systematically unfair and inefficient outcomes, defeating the very purpose, uh, on which, uh, these platforms are designed.

04:25  - Uh, and it is quite possible that a platform that is gameable, uh, may declare an impasse, uh, for an otherwise resolvable dispute. Existing research also supports this view. It shows that strategic gaming in online platforms can result in, uh, more escalation than resolution. So all of these issues set the stage for the approach that we take, uh, in our work, namely achieving immunity to gaming through strategic proof, through strategic proofness, uh, by pausing the problem, uh,

04:58  - through the lens of economic engineering, what we call as, uh, market design or mechanism design. Uh, and, and to mention one more problem with the current platforms, uh, which also motivates our study. Um, the algorithms that these systems are using, uh, are, uh, are sort of opaque algorithms. It's unknown to users. Uh, so some platforms like the cyber saddle that is used by New York City chooses not to disclose their algorithm

05:25  - because of their patent infringement, uh, concerns other platforms, famous platforms such as, uh, smart Settle, also adopt this, uh, black box, uh, approach because their algorithms involve, uh, sophisticated and unintuitive, uh, inte integer optimization techniques. So, designers of smart settle, for instance, they would openly state that their multi-issue, renegotiation system cannot be used to its full potential, uh, by inexperienced users, uh, without the assistance of a facilitator.

05:57  - So, when looked at from an economic design perspective, these issues, uh, raise red flags, you know, raise serious concerns for us, because in our area, it is widely acknowledged and already, uh, commonly accepted that, uh, the, the simplicity and the transparency of the underlying mechanics of an algorithm are very crucial for its effectiveness, for its efficacy. In economics, for example, we have this rich literature, uh, of mechanism and market design that can help, uh, to deal

06:28  - with all these issues. Uh, market design in a nutshell, uh, is a field of economics that focuses on creating mechanisms or institutions, uh, that lead to desired outcomes in a marketplace, especially when private information or strategic behavior is involved. And this literature is, is in, is based on decades of study on human incentives, and provide us, uh, with the foundations

06:53  - for designing sound institutions for practical use. So, in fact, these studies have led to several Nobel Prizes, uh, in our field, uh, based on several applications, for example, applications about auction design. How should you design, uh, an auction to sell, uh, radio spectrum? How should you design an auction to sell, uh, sponsored keywords online? How do you use it to sell oil timber? So all of these are sort of subject of our field and,

07:22  - and sound mechanisms have been developed and have been adopted, uh, in practice. How, uh, other examples are like student assignment, how students should be assigned to public schools or colleges or, uh, in, in medicine, how organ transplantation should take place. How should it be organized in the most efficient way to maximize, uh, life save? So our study is inspired by this type of literature, uh, uh, which provides us with the key set of tools

07:49  - to tackle the design of an ideal online dispute resolution platform, uh, and, and to sort of provide the objectives such as fairness, efficiency, and immunity, uh, to gaming much of the traditional mechanism design, uh, approach to bargaining, uh, which is pioneered by a seminal work of our Nobel, uh, prize winning economist, Roger Myers. And then his coauthor, uh, mark, uh, would actually equip, uh, negotiators with quasi linear, what we call quasi linear utilities.

08:18  - So you care about, uh, an an object that you receive, and also the monetary transfer, uh, that you, that you get. Uh, so, and in disputes, often money, uh, is an important issue that comes up that we see, uh, but it's rarely the only issue. So for this reason, uh, in our framework, what we do is we adopt a more general framework, uh, based on what we call an oral preference, uh, reporting approach, where, uh, we're going to ask the negotiators, uh,

08:46  - to reveal only their bargaining ranges. So, for example, tell me, uh, the range of negotiable alternatives for you, rather than giving the sophisticated information about like your utility functions, your full fledged cardinal preferences and things, sort of, that is, uh, quite abstract, uh, for the layman. Um, and these are sort of the three advantages, uh, one of the three advantages of, of our approach

09:09  - with this oral preference reporting language, we are able to, uh, develop detail free, uh, mechanisms, which is, again, hailed by Nobel Prize winning economists as a must, uh, property that needs to be satisfied for practical success. These detail free mechanisms provide, uh, robust incentives to participants, uh, by, by imposing the fundamental property of strategy proofness, uh, in our area. So an online dispute resolution, uh,

09:39  - mechanism should be strategy proof. It should be in the best interest of your participants to reveal their positions, uh, truthfully, in an honest way, without having any attempts to game the system. So this sort of sets up the backbone of our, uh, formal model, uh, where, uh, we have two negotiators, uh, who are in a dispute over two issues. Uh, for simplicity, uh, I'm going to refer to these two issues as a main issue and a supplementary issue,

10:07  - but it can be generalized to any number of issues, as long as you have more than one issue. A crucial insight, uh, in, in this work is that is the embodiment of a principle, uh, that is, that has been sort of repeatedly brought up by negotiation experts. So negotiation experts, if you take like, uh, one of the sort of bestseller books, they would say that success in mediation lies in the party's ability to expand the pie and find integrative win-win outcomes.

10:38  - Uh, so, so parties should be able to, uh, make concessions to each other. And, and this necessitates the sort of re, uh, the requirement for multi-issue negotiation and log rolling. Um, so these issues, uh, in our, uh, sort of formal model is represented by, uh, discrete and bounded sets of resolution alternatives. And each alternative represents a settlement, uh, for a particular issue. So we assume, for example,

11:07  - that each negotiator has a commonly known ranking of the alternatives, and these alternatives are diametrically opposed. So to simplify what I'm trying to say, uh, you know, suppose we're trying to divide assets. Uh, each one of us prefers to get the larger share of the assets, okay? So our preferences are diametrically opposed. You can think of it as a zero sum game, uh, in each issue. Um, but what is sort of at the crux of this matter is that

11:34  - each negotiator's bargaining range is their private information. You know, when I'm gonna walk away from negotiation, it's, it's what I know privately, and I don't have any interest in revealing that information to you or the mediator. Okay? So this information, uh, is, is sort of in economics, what we refer as the type of the agent, and you cannot force them to reveal their type. You have to, uh, design these mechanisms in a smart way,

12:02  - such that once they see the mechanism, they will realize that it's in their best interest to reveal, uh, their types, truthfully, without attempting any manipulation. Other, uh, observation in the context of dispute resolution is that depending on these type realizations, the bargaining ranges of your negotiators may not overlap. So a zone of mutual gain need not always exist, which, which means that you cannot solve every dispute. You know, there's no such thing as solving, uh,

12:31  - every dispute successfully. Um, so our hope in this study, uh, is to actually identify the ideal ODR platform. Uh, and, and in our sort of formal setting, we, we sort of abstractly say that, uh, and ODR platform is ideal if it possesses three key properties. Uh, and these are strategy proofness. Uh, the, the incentive issue that we've been talking about, it has to be immune to gaming.

13:01  - Whatever information that you are requiring your participants to reveal, they should reveal it truthfully. And the second property, as economists for us, is inevitable efficiency. Whatever resolution your platform is offering, you should not be able to find an alternative resolution that is better for both parties. Okay? So you should pick one of the optimal, uh, outcomes. And the last one, what we technically call individual rationality,

13:28  - which simply means that, uh, the offers that you are making to these negotiators should never lie outside their bargaining ranges. Whatever information they tell you, you should respect their bargaining ranges, you should not, uh, sort of jump, try to jump the gun. Um, so in, in the, uh, in our study, what we first showed, that not all disputes are necessarily solvable by an ideal, uh, by an ideal ODR protocol. So it all hinges on party's ability

13:58  - to make concessions on the issues. So in the legal terminology, for instance, this is what is referred as the quid pro quo property. So the parties should be able to trade favors, uh, and, and if the negotiators underlying preferences satisfy this quid pro quo property, which we formally sort of capture in the paper, then we show that you can actually construct an ideal mechanism and algorithm that is going to successfully solve these disputes, uh,

14:27  - meeting those three key properties. Um, so what we do in the paper is actually, uh, we provide a full characterization of these, uh, online dispute resolution protocols, uh, that are immune to gaming. IE they are strategy proof that are efficient and, and they're individually rational, providing the negotiators, uh, uh, admissible, uh, proposals. So these are, uh, what we in the paper refer

14:54  - as these log rolling mechanisms. Uh, intuitively speaking, uh, these mechanisms recommend bundles that involve compromises, uh, in the issues in the dispute. So, for example, if the res, if the resolution package that you are offering to a negotiator, uh, involves, uh, an alternative that is favorable to one negotiator, then from the other issue, you have to pick an alternative that is favorable to the opposite negotiator. So after you have to only bring these

15:22  - concession options together. Uh, and, and sort of on the mathematical side, on the sort of formal side of things, we actually show that these type of bundles, these type of log roll, what we call as log rolling bundles, have a beautiful mathematical structure. So they satisfy this mathematical property, uh, what is called the semi latice property, uh, in set theory. So any sort of those two bundles, you take two of them, you can actually construct, uh, an upper bound a bundle

15:51  - that is better than both of them, which means that you always have these external optimal members, which gives us this beautiful sort of, uh, incentive result, uh, in our characterization. Uh, and, and, and based on the structure, we actually propose a general class of, uh, resolution algorithms. And, and we show that there are multiple ways one can implement these resolution protocols, which is the kind of flexibility we would like to give these online platforms.

16:21  - You can do it simultaneously. You can offer to make them simultaneous offers. You can do it sequentially. You can approach one negotiator, tell them, you know, the, sort of the available options, or ask them to sort of provide a menu of options that is acceptable to them. And then, uh, taking these options, the, the mediator can add some options of their preference and then turn to the other negotiator

16:45  - and ask them, you know, uh, would you, would you pick one of these options? Do you find any of these options, uh, acceptable? So it can be simultaneously, it can be sequential, it can be implemented in a centralized fashion or a decentralized fashion, uh, a synchronously online, you know, in case parties are not participating in the process at the same time, which is a sort of a comfort that the online platforms are providing their users. Um, you can do a shortlisting type of, uh, resolution.

17:14  - For example, you can have one negotiator offer a shortlist of, uh, resolution options as negotiable solutions for the dispute. And the mediator would then communicate these options to the other negotiator who then, uh, would choose, uh, her favorite, uh, option from the list. So, so there is a lot of flexibility in these, uh, class of mechanisms, uh, while also ensuring that our full deserter of strategy, proofness efficiency and individual rationality, uh, is met.

17:44  - Remember, we also assume that, uh, the underlying preferences are compatible with the squid pro quo assumption that the, the negotiators are able to, uh, make, make concessions among themselves. And, and this allows us to always, uh, make a recommendation from the set of, uh, log rolling bundles. The, the set, uh, consists of optimal elements. So this is not something we would like to sort of, uh,

18:08  - of course, uh, implement in practice, but from the sort of mathematical beauty of the theory, you can actually, uh, obtain if you, if you wanted to favor one side of the dispute, you can actually obtain, uh, the sort of algorithms that would, uh, favor one side of the dispute while picking, you know, those concessional, uh, outcomes that is in the best interest of one side in a systematic way. Uh, but of course, we don't do that. We don't wanna do that as, and, and,

18:36  - and we focus in the study at the, uh, central element of our class of mechanisms, which we refer as a, uh, constrained, shortlisting, uh, mechanism. And we show that this actually satisfies a very natural fairness requirement. If you insist on fairness, then you should always use the central element, which is sort of anonymous, which doesn't favor each size. It is unbiased and, and can be sort of very easily implemented in practice.

19:03  - So that sort of forms the fundamental, uh, proposal of the paper. I would say that as online dispute resolution, uh, systems become more prevalent around the world, uh, mediation, the traditional mediation is being integrated into online platforms where disputes, particularly those arising from e-commerce or other digital tra transactions, as well as offline disputes, are resolved through, uh, digital means.

19:30  - And the, the idea of automating, uh, the negotiation process, simplifies decision making and speeds up the resolution process. So it it, but to do this, you need robust algorithms to ensure that, uh, strategic manipulation, you know, the, the ga, the non-game ability of the system, uh, doesn't undermine the intended fairness and efficiency, uh, of the, of the resolution outcome. Remember the incentives property. If you don't have good incentives, then no matter you know,

20:00  - how good your algorithm is in terms of efficiency and fairness, people are going to distort it, and the outcome is not going to be fair or efficient. That's why we argue that a market design approach where the principles of economics has to be sort of at the heart of this engineering problem, uh, is necessary to improve the efficiency and fairness of, uh, existing online dispute, uh, resolution, uh, problems. Uh, and the, what is the implication of this?

20:28  - Well, um, the rise of online dispute resolution has major implications for legal systems, uh, worldwide, in particularly in terms of reducing court backlogs, increased cost, saving time-wise, financially and efficiency. And governments and courts must consider how to integrate these systems, uh, while ensuring transparency and fairness, uh, as well as, uh, good incentive properties. So understanding incentives is extremely important, uh, for, uh, having the ultimately desirable, uh, outcomes, uh,

21:01  - in online dispute resolution, uh, processes, uh, for, for future. So our, our studies is, I believe, uh, in market design in our field is the first to tackle this problem through, uh, economic design, uh, perspectives. Uh, but for future, there's more research that is necessary for designing robust mechanisms that resist gaming, particularly in situations where the parties are going to have as symmetric information or as symmetric experience.

21:29  - So, future studies should focus on ensuring that all participants, regardless of their background or experience, are incentivized to participate truthfully, uh, giving the information, uh, that is, that is truthful and honest. So, given that many ODR systems, uh, the currently existing systems are difficult for inexperienced people to navigate without additional help, research, uh, should focus on making these systems more intuitive and user-friendly without compromising on the

21:59  - complexity, uh, needed. For fairness to sum up, uh, designing successful online dispute resolution platforms is a new research area. And I do sincerely hope that the industry practitioners would involve more economic designers to help, uh, build platforms that are simple, more efficient, fair, transparent, and that give their participants the right incentives to act.

22:25  - Uh, truthfully, this should help increase the number of disputes successfully, result, uh, and quickly result while, uh, boosting the user satisfaction, uh, and ultimately, uh, making the world a happier place.

Faculti AI can make mistakes