The landscape of dispute resolution has undergone a profound transformation, driven by technological advancements that were significantly accelerated by the COVID-19 pandemic. In response, two pivotal sets of guidelines have emerged in 2025 to steer the ethical integration of technology and artificial intelligence (AI): the 2025 AFCC/ABA Model Standards for Family and Divorce Mediation and the ICODR/NCTDR Standards and more recent Guidance for Third Parties Using Artificial Intelligence. While the AFCC standards provide a broad, human-centric ethical framework specifically for family law, the ICODR guidance offers a granular, tech-specific deep dive into the unique risks and requirements of AI platforms.
Scope and Philosophical Foundations
The 2025 AFCC Model Standards, the result of a three-year collaborative effort between the AFCC and the ABA Section of Family Law, are designed as the “ethical cornerstone” for family and divorce mediation. These standards are participant-centered, grounded in values of integrity, fairness, and safety, with a specific emphasis on protecting children and survivors of domestic abuse.
The 2025 revision was necessary to address issues omitted in previous versions, most notably the use of technology and online dispute resolution (ODR). AFCC defines technology broadly as an evolving array of digital tools—including AI, large language models (LLM), and algorithmic decision-making—intended to enhance, but not define, the process.
In contrast, the ICODR AI Guidance is process-agnostic, but tech-specific, applying to all forms of “conflict engagement,” including mediation, arbitration, and coaching. Its primary mission is to bridge the “time lag” between rapid AI innovation and slower-moving professional standards.
While AFCC views AI through the lens of family-specific needs like “the best interests of the child,” ICODR views AI through the lens of ODR principles such as accessibility, accountability, and transparency.
The Role of the Human: Oversight and Impartiality
A critical point of comparison is the degree of human oversight required. The 2025 AFCC standards emphasize that the mediator remains an impartial facilitator who must not substitute for independent legal advice or therapy. Standard IX mandates that technology should serve as “a tool,” and mediators must remain attentive to how it impacts “communication dynamics” and “participant engagement.”
The 2025 AFCC Standards use specific levels of guidance: “shall” for mandatory practices, “should” for highly desirable ones, and “may” for those left to professional judgment.
The ICODR guidance goes much further in its technical requirements for oversight. It explicitly requires “human oversight” of AI-based interventions throughout the entire process—including administrative support, analysis, recommendations, and decisions. ICODR introduces the concept of “traceability,” demanding that the “path to outcome” be auditable when AI is employed. Furthermore, ICODR warns against “authority bias,” where humans might unconsciously grant AI-generated outputs “perceived normative or evidentiary weight” rather than exercising critical assessment.
Informed Decision-Making vs. AI Transparency
Both the AFCC and ICODR Standards and Guidance converge on the necessity of informed consent, but they define the “informed” aspect with different levels of technicality. For AFCC, Standard II: Informed Decision-Making focuses on ensuring parties have access to sufficient “information and knowledge” to make voluntary choices. Under Standard IX, mediators must provide clear information about the technology’s “benefits and challenges” before obtaining consent.
ICODR’s “Transparent” standard requires a deeper “look behind the curtain.” Practitioners must disclose the “magnitude of technology’s influence” on restricting or generating options. ICODR highlights a fundamental difference between traditional digital communication (like email) and AI: whereas traditional tech delivers data exactly as created, AI generates “synthetic information” that inherently reflects the biases and limitations of its training data. Consequently, ICODR requires practitioners to disclose the sources and methods used to gather data that influences any AI-driven decision.
Addressing Bias and Power Imbalances
In the realm of family mediation, power disparities are a primary concern. The AFCC standards mandate rigorous, separate, and confidential screening for domestic abuse and child maltreatment. Mediators must assess whether technology might “enable further harm” by allowing abusers to continue patterns of intimidation or coercion through digital means. The 2025 standards shifted to more inclusive terminology, moving from “domestic violence” to “domestic abuse” to encompass coercive control.
ICODR addresses power imbalances through its “Equal” standard, which strives to ensure that “offline privileges and disadvantages are not replicated” in the digital process. It calls for proactive efforts to prevent AI from “creating, replicating, or compounding bias.” ICODR prompts mediators to consider the “ecological cost” of AI and whether “unequal access” to sophisticated AI tools might provide one party with a technological or informational advantage.
Confidentiality, Security, and Data Integrity
The 2025 AFCC standards require mediators to implement “robust data security measures,” such as end-to-end encryption, and to treat digital data with the same level of confidentiality as traditional mediation. Standard VIII also warns mediators about the limitations of digital confidentiality, especially regarding subpoenas for “digital or electronically stored information.”
ICODR expands this by requiring practitioners to understand the specific data use policies of AI platforms. It asks critical questions: “Is the party data used to train the AI’s future models” and “How long is the information retained before deletion?” ICODR emphasizes that the human user retains responsibility for ensuring that any AI-influenced decisions are accurate and lawful.
Competence and Professional Training
The AFCC’s updated Standard XII significantly expands requirements for professional competence. Mediators are now expected to have specific skills in “conducting online mediation” and understanding the “responsible and ethical use of technology.” This includes being aware of “privacy laws” and the “potential impact of technology on communication.”
ICODR mirrors this, but adds a layer of technical diligence. Competence for ICODR includes the ability to explain the AI applications to parties and counsel. Practitioners should exercise “due diligence” when relying on AI outputs and must maintain “independent verification” of information that influences procedural fairness. Interestingly, ICODR also suggests that mediators should be aware of “legal explainability”— the lack of which may obscure systemic biases and jeopardize the trustworthiness of the process.
Conclusion: A Complementary Relationship
The AFCC 2025 Model Standards and the ICODR AI Guidance are not redundant; rather, they are complementary. The AFCC standards provide the broad ethical boundaries necessary for the sensitive, human-centric nature of family reorganization, ensuring that self-determination and safety are never sacrificed for efficiency. The ICODR Standards and Guidance provides the technical guardrails for the “how” of AI integration, offering a checklist of questions that force practitioners to reckon with the “synthetic” and potentially biased nature of AI outputs.
On September 14, 2001 President Bush stood on the ruins and ashes of Ground Zero. He was addressing the rescue workers wearing a sports jacket with an open-collar shirt. The...
By Noa ZanolliThe murder of UnitedHealthcare CEO Brian Thompson has triggered polarized reactions, reflecting deeper frustrations with the healthcare claims and dispute systems in the US. Allegations of systemic barriers—popularly encapsulated in...
By Seun Lari-WilliamsINTRODUCTION Many South African Schools are facing demands to change and even to radically transform themselves. Some of these demands relate to the ‘isms’ such as racism, and other identity...
By Felicity Steadman, John Brand
