AI and online mediation reshape civil justice

In the field of civil justice, the integration of alternative dispute resolution (ADR) and mediation into court processes has developed in response to the limitations of formal adjudication. As Sir Geoffrey Vos observed in 2021, ADR is in no sense “alternative”, but should form a mainstream element of the civil justice process. It has been associated with affordability, flexibility, efficiency, and a more accessible, user-focused approach to resolving disputes. More recently, the growth of digital technology and artificial intelligence (AI), particularly as a tool for enhancing communication, has increased interest in online dispute resolution (ODR), where disputes are resolved through digital platforms. The concept of ODR is broad and can encompass many of the principles that underpin ADR and mediation.

Lon Fuller’s observation in 1971 that mediation is “all process and no structure” remains influential. Mediation is defined less by rigid rules than by its core aim: to enable parties to resolve their dispute through facilitated dialogue. The mediator draws on skill and experience to support constructive, forward-looking engagement between the parties. However, traditional definitions of mediation as being voluntary, impartial and confidential are increasingly open to challenge. For example, following the recent case of Churchill v Merthyr Tydfil County Borough Council, the notion that mediation must be voluntary has been qualified, as courts may now order parties to engage in ADR. At the same time, as AI becomes more sophisticated and ODR more widespread, mediation is evolving in response to technological change. While these developments offer opportunities to expand access and efficiency, they also place pressure on core features of the process, particularly neutrality and confidentiality. 

The idea of technology acting as a “fourth party” in dispute resolution is not new. Early examples demonstrate how digital systems can support negotiation and mediation at scale. The marketplace platform eBay, for example, developed an ODR system capable of resolving millions of disputes annually, initially drawing on mediation principles before evolving into a more automated, data-driven process. Similarly, the Civil Resolution Tribunal in British Columbia, Canada, employs structured, decision-tree tools to guide users through early-stage dispute resolution, offering tailored pathways for low-value claims. Other platforms, including Smartsettle and Cybersettle, use predictive analytics to identify areas of agreement and generate settlement options. These systems illustrate how AI can support dispute resolution by structuring negotiation and identifying and suggesting mutually acceptable outcomes. 

The capacity of AI to analyse large volumes of data, generate text and identify patterns offers a further dynamic for those involved in mediation. AI can assist mediators by summarising documents, highlighting key issues, and supporting preparation for mediation and negotiation. Its predictive capabilities may be used in forms of early neutral evaluation or to assist mediators in managing expectations and facilitating “reality-testing” during discussions. At a practical level, digital platforms such as Zoom or Teams enhance flexibility, enabling mediators to operate online and convene joint sessions or private discussions through virtual break-out rooms. These developments can improve efficiency and accessibility, particularly in long-distance, complex or document-heavy disputes. For the individual mediator, the option of using online communication platforms, allows for greater flexibility, convenience and efficiency in organising and running a mediation. 

Notwithstanding these benefits, the use of AI in mediation raises important questions of procedural fairness and justice. Mediation depends on the impartiality and neutrality of the mediator. While human neutrality is not absolute, it is grounded in professional standards and ethical obligations. The CEDR definition of a mediator, for example, found in its Model Procedure, requires the mediator to be neutral and impartial. AI systems, by contrast, cannot be assumed to be neutral. Their outputs are shaped by underlying data and design choices, and their operation is often opaque. The “black box” nature of many AI tools makes it difficult to scrutinise how conclusions are reached, creating risks of over-reliance on outputs that may not be fully understood or interrogated. This has implications for the integrity of the mediation process and for the confidence of those who participate in it. 

Online mediation also raises distinct concerns in relation to transparency and confidentiality. Despite confidentiality provisions cemented into mediation agreements, parties may easily record sessions without consent in an online environment. The availability of AI-powered transcription tools further complicates matters, as conversations may be automatically recorded and stored. In addition, documents uploaded to AI systems may be retained or used in ways that are not transparent to users. The input of sensitive or legally privileged information into public AI platforms presents risks, potentially compromising confidentiality. Concerns about privilege have also been highlighted, with questions arising as to whether information shared with AI systems can retain its protected status.  These issues were flagged in a speech recently by Sir Colin Birss, President of the High Court, and are particularly relevant for mediators. Parties may input the terms of a settlement agreement into a generative AI tool and potentially breach confidentiality, for example.   

In this context, the mediator’s role could become more complex when they are operating in an online environment. In an online mediation it is not enough to focus on the interpersonal dynamics of the dispute; there is also a need to manage the technological environment in which the process takes place. Mediators must take active steps to ensure that online platforms are secure, that participants understand the risks associated with AI tools, and that appropriate safeguards are in place to protect confidentiality. This requires not only understanding and technical awareness but also the development of new professional norms and guidance, such as those created by the IBA Mediation Committee, as an example.

The integration of ADR into civil justice has always been driven by the objective of delivering more affordable, efficient and creative forms of dispute resolution, while maintaining its emphasis on collaboration and constructive engagement. AI has the potential to advance these aims, supporting mediators in both traditional and innovative settings. The continued growth of ODR provides an opportunity to rethink how mediation can evolve in a digital environment. However, its use also introduces ethical challenges. In particular, the impact of AI on neutrality, confidentiality and fairness requires careful consideration.

The rapid development of AI provides an opportunity to rethink how mediation can evolve in a digital environment. The challenge is to ensure that technological innovation is aligned with the fundamental values underpinning the process. This will require greater focus on standards, increased digital education and literacy among mediators, and a commitment to maintaining human oversight in decision-making. The growth of ODR presents a valuable opportunity to rethink how ADR and mediation can evolve, combining definitional disruption and technological innovation with strong ethical foundations, to shape a more accessible and effective future for dispute resolution.