With the rapid growth of ODR in a number of sectors it is vital to establish ethical standards to undergird the design, structure, practices, and implementation of global online dispute resolution systems. This simplified standard set builds on previous work by the National Center for Technology and Dispute Resolution on principles for ODR practice. Taken together they are intended to provide a touchstone for best practices, rules, qualifications, and certification efforts for online dispute resolution processes and practices.
These ODR Standards apply to ODR practitioners and to technological platforms, systems, and tools when employed for dispute handling. They are interdependent and must be applied together. They can be useful for ODR software and system developers and to inform the public of requirements for ethical, technology-infused dispute resolution. Reference to “ODR” in the Standards includes people, entities, and technologies involved in implementing, hosting, or providing ODR services.
ODR Standards require that online dispute resolution platforms and processes must be:
1. Accessible
ODR must be easy for parties to find within a system and participate in and not limit their right to representation. ODR should be available in communication channels accessible to all the parties, minimize costs to participants, and be easily accessed by people with different types of abilities.
2. Accountable
ODR systems must be continuously accountable to the institutions, legal frameworks, and communities that they serve. ODR platforms must be auditable and the audit made available to users. This must include human oversight of: i) traceability of the originality of documents and of the path to outcome when artificial intelligence is employed, ii) determination of the relative control given to human and artificial decision-making strategies, iii) outcomes, and iv) the process of ensuring availability of outcomes to the parties.
3. Competent
ODR providers must have the relevant expertise in dispute resolution, legal, technical execution, language, and culture required to deliver competent, effective services in their target areas. ODR services must be timely and use participant time efficiently.
4. Confidential
ODR providers must make every genuine and reasonable effort to maintain the confidentiality of party communications in line with policies that must be articulated to the parties regarding i) who will see what data, ii) how and to what purposes that data can be used, iii) how data will be stored, iv) if, how, and when data will be destroyed or modified, and v) how disclosures of breaches will be communicated and the steps that will be taken to prevent reoccurrence.
5. Equal
ODR providers must treat all participants with respect and dignity. ODR must seek to enable often silenced or marginalized voices to be heard and strive to ensure that offline privileges and disadvantages are not replicated in the ODR process. ODR must provide access to process instructions, security, confidentiality, and data control to all parties. ODR must strive to ensure on an on-going basis that no process or technology incorporated into ODR provides any party with a technological or informational advantage due to its use of ODR. Bias must be proactively avoided in all processes, contexts, and regarding party characteristics. ODR system design must include proactive efforts to prevent any artificial intelligence decision-making function from creating, replicating, or compounding bias in process or outcome. Human oversight is required in ODR system design and auditing to identify bias, make findings transparent to ODR providers and users, and eliminate bias in ODR processes and outcomes.
6. Fair and Impartial
ODR must treat all parties equitably and with due process, without bias or benefits for or against individuals, groups, or entities. Conflicts of interest of providers, participants, and system administrators must be disclosed in advance of commencement of ODR services. The obligation to disclose such circumstances shall be a continuing obligation throughout the ODR process.
7. Legal
ODR providers must abide by, uphold, and disclose to the parties relevant laws and regulations under which the process falls
8. Secure
ODR providers must make every genuine and reasonable effort to ensure that ODR platforms are secure and data collected and communications between those engaged in ODR are not shared with any unauthorized parties. Disclosures of breaches must be communicated along with the steps taken to prevent reoccurrence.
9. Transparent
ODR providers must explicitly disclose in advance and in a meaningful and accessible manner: i) the form and enforceability of dispute resolution processes and outcomes and ii) the risks, costs including for whom, and benefits of participation. Data in ODR must be gathered, managed, and presented in ways to ensure it is not misrepresented or out of context. The sources and methods used to gather any data that influences any decision made by artificial intelligence must be disclosed to all parties. ODR that uses artificial intelligence must publicly affirm compliance with jurisdictionally relevant legislation, regulations, or in their absence, guidelines on transparency and fairness of artificial intelligence systems. ODR must clearly disclose the role and magnitude of technology’s influence on restricting or generating options and in final decisions or outcomes. Audits of ODR systems and platforms must identify metrics used to assess system performance, making the accuracy and precision of these metrics known and accessible to any ODR system operator and user. Users must be informed in a timely and accessible manner of any data breach and the steps taken to prevent reoccurrence.
IMPLEMENTATION OF ODR STANDARDS
We recommend practitioners, platform and system designers, and managers of any form of online dispute resolution process commit to these ODR Standards. We encourage dispute resolution membership organizations and other entities–whether public or private–with responsibility for and/or authority over ODR-related processes, practices, and practitioners to incorporate these standards and create appropriate mechanisms for accountability and monitoring including through the examination of impacts of different technological designs and use across disputant demographics and case types. © 2017 and May 2022, National Center for Technology and Dispute Resolution and International Council for Online Dispute Resolution.
1) Mediator remains accountable.
AI may assist; it does not replace mediator judgment or responsibility.
2) No undisclosed substantive influence.
If AI is used to analyze options, generate proposals, evaluate likely outcomes, or shape substantive direction, that is substantive assistance and should be disclosed.
3) Data minimization + confidentiality by design.
Use the least identifying info necessary; prefer local/offline or enterprise-secured systems when feasible; define retention/deletion rules.
4) Neutrality and equal access.
If one side receives AI-driven analytic advantages through the mediator, equivalent access (or functional equivalence) should be provided to the other side.
5) No “authority laundering.”
AI outputs must never be presented as “what the model says is fair” or “the right answer.” Output
should be framed as drafting/brainstorming material.
6) Explainability at the right level.
Parties should understand what AI is doing (summarizing vs generating options vs
evaluating) without requiring technical detail.
Standard 1: Purpose and Scope
A mediator shall conduct mediation in a manner that supports (a) participant
empowerment and (b) problem-solving optimization, consistent with party self-determination and mediator impartiality.
Empowerment means supporting parties’ capacity to understand, choose, and act
voluntarily.
Optimization means supporting processes that increase decision clarity, option quality, and agreement durability—without directing outcomes.
Standard 2: Empowerment Practices:
A mediator should employ practices that enhance participant capacity, including:
• Supporting informed participation (process clarity, role clarity, decision points).
• Encouraging articulation of interests, priorities, constraints, and concerns.
• Ensuring opportunities for questions, reflection, and private consultation.
• Actively monitoring imbalance that impairs meaningful participation and using process adjustments to restore effective voice.
Standard 3: Optimization Practices
A mediator should support a problem-solving process that improves solution quality, including:
• Option-space expansion before convergence (multiple pathways explored).
• Tradeoff clarification (priorities, risks, timing, values, feasibility).
• Reality-testing of proposals (workability, implementation steps, foreseeable obstacles).
• Durability design (specificity, milestones, follow-up methods, revision triggers).
Standard 4: Non-Directive Optimization
A mediator shall not use “optimization” to pressure agreement or to steer parties toward an outcome preferred by the mediator or any third party.
Optimization methods must remain: Process-focused (improving thinking and comparing), and Choice-preserving (parties decide).
Standard 5: Ethical Use of AI for Empowerment and Optimization
A mediator may use AI tools to support empowerment and optimization only when such use is consistent with confidentiality, party self-determination, fairness and transparency. A mediator shall:
• Disclose and obtain consent for substantive AI uses (e.g., generating options, analyzing proposals, summarizing party statements, drafting terms from negotiation content).
• Protect confidentiality through data minimization and appropriate security practices.
• Avoid presenting AI outputs as authoritative, determinative, or “the fair answer.”
• Ensure that AI assistance does not create unfair informational advantage between participants.
Standard 6: Transparency and Record Practices
A mediator shall clearly describe, in plain language, the nature of AI use, the categories of data involved, and retention/deletion practices.
Where feasible, the mediator should provide:
• A brief AI-use disclosure statement (administrative vs substantive)
• A retention/deletion commitment
• An opt-out path and alternative process
Standard 7: Competence
A mediator who uses optimization methods or AI tools shall maintain competence in their responsible use, including:
• Understanding limitations, bias risks, confidentiality risks
• Reviewing AI outputs before use and avoiding over-reliance
Standard 8: Continuous Improvement A mediator should periodically evaluate whether empowerment and optimization practices are improving party experience and agreement durability, and should adjust processes accordingly.
Amendment to Self-Determination Standard – Add subsection:
A mediator should employ practices that enhance participant capacity to make informed, voluntary, and effective decisions, including clarification of interests, exploration of options, and evaluation of tradeoffs, while preserving party control over outcomes.
Amendment to Quality of the Process / Competence Standard – Add subsection:
A mediator should support a problem-solving process that improves the clarity, feasibility, and durability of agreements by encouraging option-space expansion, reality-testing, and implementation planning.
Amendment to Impartiality Standard – Add subsection:
The use of structured problem-solving techniques or optimization tools shall remain non-directive and shall not be employed to steer parties toward particular outcomes.
Amendment to Confidentiality Standard (AI-Responsive) – Add subsection:
Where technological or AI-assisted tools are used to support substantive aspects of mediation, the mediator shall ensure appropriate data protection, minimize identifiable information, and disclose data handling practices.
New Section: Use of AI & Decision-Support Tools in Mediation
A mediator may use analytical or AI-assisted tools to support participant empowerment and problem-solving optimization provided that the mediator:
a) Discloses and obtains consent for substantive uses; b) Maintains confidentiality and data security; c) Avoids presenting outputs as authoritative or determinative; and d) Ensures fairness and neutrality in access and influence.
Amendment to Continuing Education / Competence – Add subsection:
Mediators employing advanced problem-solving methodologies or technological tools should maintain ongoing education regarding their ethical, practical, and procedural implications.
Divorces often start from a place of conflict or stress, but some divorces are more contentious than others. A number of circumstances factor into a high-conflict divorce, and each one...
By Sarah JacobsIntroduction The concept of law is as ancient as mankind. From domestic interaction to group structures, rules of law have impacted how individuals deal with conflict. Rules that govern human...
By Alan SimpsonI've directed my readers to Adir Waldman's fine book Arbitrating Armed Conflict before. Now that there is pitched battle in the Middle East with significant civilian casualties, I once again...
By Victoria Pynchon
