Is Kleros a Fair Dispute Resolution System?

By Federico Ast and Daniel Dimov

Kleros' approach to arbitration is radically different to traditional court systems and alternative dispute resolution methods. Relying on cryptoeconomics, it provides crowdsourced jurors the incentive to arbitrate on various types of disputes.

The goal is to bring fast, affordable and secure resolution to a number of disputes native of the Internet Age.

Kleros promises to transform the field of dispute resolution in a way similar to how Wikipedia revolutionized encyclopedia publishing. Back then, few people expected that an encyclopedia written by anonymous people on the Internet could compete with traditional encyclopedias. Nevertheless, Wikipedia ultimately triumphed and forced Microsoft to shut down its paid product Encarta. It also brought the end to the venerable Britannica, 242 years after the first edition.

A Chinese edition of Encyclopedia Britannica.

Sceptics believe that Kleros does not provide a fair method for dispute resolution and that this will become an obstacle for mass adoption. Our goal in this article is to address this concern: can Kleros resolve disputes in a fair manner?

Assessing Fairness in Crowdsourced Online Dispute Resolution (CODR)

In order to assess the fairness of Kleros, first we need to identify reliable fairness criteria. Daniel Dimov (one of the coauthors of this article) has worked extensively in the development of a model for assessing procedural fairness in the context of CODR (read his PhD dissertation on fairness in CODR methods).

Procedural fairness refers to a set of criteria for determining whether a given procedure contributes to a fair outcome. Dimov examines two types of procedural fairness, namely, objective procedural fairness and subjective procedural fairness, and merges them in a combined framework.

Objective procedural fairness is defined as the extent to which the procedure complies with the fairness standards defined by the Directive on Alternative Dispute Resolution (ADR), an EU law applying to the entire European Union.

Subjective procedural fairness is defined as: “an individual’s subjective perception of the fairness of a procedure”. This is based on empirical studies about people’s subjective perceptions regarding various aspects of procedures. It falls under the domain of psychology as it relates to human behaviour and mental processes (and not to objective normative frameworks).

The combination of objective and subjective procedural fairness is important to assess the overall fairness of a dispute resolution procedure. A procedure may be legally fair, but subjectively regarded as unfair by the disputants.

Let’s see how Kleros scores with regard to the criteria of Dimov’s framework of procedural fairness.

Evaluating the Procedural Fairness of Kleros on the Basis of Dimov’s Framework

1) Expertise

The third neutral party should have 1) the necessary knowledge and skills in the field of alternative or judicial resolution of consumer disputes and 2) a general understanding of the law.

Kleros jurors self-select into the subcourt where they wish to conduct arbitration. Kleros does not ask for the jurors' real identity or to prove they are qualified to arbitrate disputes in the subcourt where they want to work.

The expertise requirement is conducted via economic incentives. Kleros generates for users the incentive to self-select for the subcourts where they have expertise. Users who self-select into the courts for which they have the right skills will, on average, make money over time. Users who self-select into courts where they don't have the right skills will lose money and tend to abandon the system.

Even though, in theory, jurors may not have subject matter expertise (anyone can participate in the subcourt), in practice, users without adequate expertise would suffer an economic loss and exit the subcourt (unless they wish to lose money while they work, in order to gain those skills). This works similarly to Wikipedia in the sense that a user who does not have expertise in a field to which an article edited by him/her relates, may still edit the article but will likely be sanctioned by Wikipedia.

2) Independence

Jurors should have personal and functional independence. Personal independence refers to guarantees that are built into the position of third neutral party regarding the duration of appointment, pay and dismissal. Functional independence means that the third neutral party should feel free to freely give his judgements.

Smart contracts guarantee personal independence in the sense that no authority could affect the duration of the appointment (jurors are drawn for the whole court session), pay (arbitration fees are written in the smart contract) and dismissal (there is no way to dismiss a juror before the end of the court session).

Pseudonymity favors functional independence. The fact that jurors do not need to reveal their true identity makes them less vulnerable to attacks or intimidation. However, token redistribution (which gives jurors the incentive to vote in coherence with the final decision) lowers functional independence. Functional independence may be retrieved by forking, in case the results would be significantly different from what users consider as a fair outcome.

3) Impartiality

The third neutral party must not have any internal prejudices, prejudgments or predisposition toward some parties or some of the elements of the subject matter of the dispute.

The procedure for random selection of jurors among those who staked tokens in a subcourt makes it hard to attack the court.

It is extremely hard for a juror to be able to be drawn into a court where he has a vested interest (to learn more about how this works, read this white paper). In practice, it would be extremely unlikely for jurors having a vested interest in a case to compose a significant part of the drawn jurors.

It is possible that internal biases still exist in jurors. There are some ways within Kleros to present information in such a way that it minimizes this bias. However, it could be also argued that no system can be completely free of biases. For an in depth discussion about this, read this article.

4) Transparency

The dispute resolution process should be understandable and, if necessary, possible to replicate.

Kleros’ dispute resolution procedure is documented in many places. All cryptoeconomic research is public and the code is open source. A fully working version of Kleros could be replicated in matter of minutes by anyone with technical skills in blockchain.

5) Fair hearing

Parties should be provided with 1) a notice informing them about the commencement of the dispute resolution process and 2) an opportunity to present their cases and rebut the cases of their opponents.

The start of a dispute is public. Any party having set a way to get informed (this can be done by just providing an email address) can be informed about the commencement of a dispute. Parties have the opportunity to present arguments on their defense.

6) Counterpoise

Pre-existing imbalances in the financial status of parties and the computer skills of the disputants should be neutralized.

As Kleros early use cases are connected to disputes within the blockchain ecosystem, arguably there are not high imbalances in computer skills.

Kleros could make a better job of balancing financial inequalities than competitors such as customer support services. The customer support service of an Internet platform such as eBay is more likely to favor the wealthier party, as it is the one generating more revenue for the platform. Furthermore, customer support services are usually provided in a limited number of languages. This puts foreign language speakers at a disadvantage compared to native language speakers.

Imbalances in the financial status of parties could lead to inequality in the appeal process. Parties with lower financial resources will have less opportunity to appeal (as the losing party needs to pay for appealing). The post-dispute insurance process is designed to neutralize this imbalance.

7) Ensuring a reasonable length of procedure

The outcome of the dispute resolution procedure should be available within a period of 90 calendar days from the date on which the provider of the dispute resolution service has received the complete complaint file.

While the length of procedure in Kleros would depend on the type of dispute, it would be arguably faster than many other alternative dispute resolution methods. The reason is that the amount of users from which jurors are pooled is much higher, which generates a large supply of arbitration services.

In theory, the number of appeals could lead to a lengthy procedure. But, in practice, frivolous appeals in order to delay a final ruling are expensive and lead to a compensation that will be paid to the victim of those appeals.

8) Providing reasons

Jurors must give parties a statement of the grounds on which the decision is based.

In Kleros, jurors are required to provide a short text explaining their vote.

9) Voluntary participation

The participation of parties in the dispute resolution procedure should be voluntary.

As in traditional arbitration, parties have to agree on Kleros as dispute resolution method before the dispute happens.

10) Process control

The increase of the control over the development and selection of information that will constitute the basis for making a decision strengthens the perceptions of procedural fairness.

The information that jurors will evaluate for making a decision depends on the type of case and the subcourt. When parties agree to a specific subcourt to settle a dispute, they also agree to the type of evidence that will be used for resolution.

11) Decision Control

Decision control refers to the extent to which the parties are free to reject or accept a decision rendered by a third party. The increase of the extent to which the participants in a procedure can reject or accept a decision rendered by that procedure strengthens the perceptions of procedural fairness.

Parties can appeal to decisions they disagree with and have their case ruled again. This gives them control over the decision.

12) Consistency

The consistent application of a procedure across persons and across time strengthens the perceptions of procedural fairness.

Kleros fully complies with this condition. All cases within a subcourt are settled with the same subcourt policy. The procedure maintains the consistency across time, as any change to the subcourt policy needs to be approved by users through the governance mechanism.

13) Accuracy

Individuals view procedures based on accurate information as more fair than the procedures based on inaccurate information.

Information used in the resolution process is agreed beforehand by the parties. Blockchain immutability guarantees that parties cannot tamper with the evidence.

14) Correctability

The opportunity to correct a decision strengthens the perceptions of fairness of the procedure used for making the decision.

Kleros allows for participants to appeal decisions that they consider unfair. Such decisions will be reviewed again by a new jury that will consist of twice as many jurors plus one.

15) Ethicality

The increase of the extent to which a procedure conforms to personal standards of ethics and morality increases the extent of the perceived fairness of the procedure.

Using Kleros as arbitrator is a choice. Parties with different standards of ethics and morality are not required to use Kleros. Besides, the creators of different subcourt policies will be able to implement in their subcourt policies ethical principles to which most disputants will adhere, thus ensuring a high-level of ethicality.

Is Kleros Fair?

Kleros seems to comply with most of the points of Dimov’s framework for fairness in online dispute resolution systems. Furthermore, the fairness of Kleros is enhanced by the use of blockchain technology. Most other CODR procedures are centralized. This means that the operator may interfere in the procedure and breach Dimov’s fairness framework.

This may have happened when, on the 31st of January of 2012, eBay decided to stop its CODR procedure called Community Review Forum. In centralized systems, the owner can decide to stop the procedure at any time and for any reason. As Kleros replaces the operator of the CODR procedure with a decentralized autonomous organization, it removes the centralized trusted operator and gurantees a high level of procedural fairness.

Dr. Federico Ast. Co-Founder and CEO at Kleros.

Dr. Daniel Dimov. Lawyer specialized in Internet law. PhD (Leiden University) in the field of crowdsourced online dispute resolution.

Join Kleros!

Join the community chat on Telegram.

Visit our website.

Follow us on Twitter.

Join our Slack for developer conversations.

Contribute on Github.