Artificial intelligence (AI) for helping mediate disputes has begun to gain traction. Although it might not yet be commonplace, it is on its way to playing a pivotal role in conflict resolution, providing mediators with tools and functionalities that streamline the mediation process, offer data-driven insights, and help manage communications and negotiations. The use of AI in mediation looks both promising and transformative and presents opportunities and challenges for mediators.
The Transformative Impact of Artificial Intelligence on the Mediation Process
One can’t help but think of artificial intelligence (AI) without conjuring up Hollywood-inspired images of computers exhibiting human-like qualities in their quest for control. While these images remain mostly the stuff of imagination, the exponential growth of machine learning, predictive algorithms and their implications for the practice of law and mediation are too real to ignore. Clearly, the “AI revolution” has arrived and its transformative impact is here to stay.
Yet, most mediators still question whether its arrival portends the replacement of the genuinely human skills that mediators bring to the table. Creativity, empathy, and emotional intelligence are a few of the so-called “soft skills” that remain the exclusive province of human intervention. That said, I recently attended several international mediation conferences where the subject of artificial intelligence dominated agendas, and panel discussions moved beyond a dystopian future vision and focused instead on the practical benefits of having this fourth party in the mediation room.
What will be the practical impact of technological advancements like artificial intelligence in preparation for and during the mediation process? What guardrails exist to ensure the ethical and unbiased application of machine learning in a mediation environment? These are but a few of the questions that we, as mediators, must ask ourselves to keep pace with the changing world and its new methods for resolving disputes.
Client Expectations
As attorneys and clients come to mediation from a world increasingly accustomed to using artificial intelligence, one can assume that they will expect no less from their chosen mediator. Online dispute resolution, while technologically advanced, may not fully replace the human touch in mediation. What might this brave new world of artificial intelligence in mediation look like, and what should the mediators of tomorrow be thinking about today?
Using Artificial Intelligence as “Augmented Intelligence”
Artificial intelligence augments and enhances mediators’ capabilities through advanced data processing, predictive analytics, natural language processing, and various support tools. This augmentation allows mediators to make more informed decisions, manage cases more efficiently, and ultimately improve the mediation process. Some of the ways are discussed below.
Preparation
AI can assist as a thought planner during the convening phase of mediation. The mediator can use AI tools to identify objectives and priorities, preview potential obstacles, and even attempt to gauge what might be a “fair resolution.” Questions can be scripted in advance based on the specific dispute, such as “Give me five examples of interests that might be important to the parties” or “Provide examples of how the competing interests of the parties might be reconciled.”
For mediators who routinely receive large volumes of expert reports, business records, or other documents, AI-based software can transform the exercise from one of locating and synthesizing information to an evaluation and verification process.
Mediation Brief Analysis
Artificial intelligence can help summarize briefs, analyze and evaluate the strength of legal positions, and even offer predictions about an attorney’s advocacy skills and mediation competency based solely on the composition of the mediation brief. At Signature Resolution, we are actively experimenting with integrating AI to enhance our case data through AI-generated summaries and advanced search capabilities.
Communication
For those interested in a deeper level of mediation preparation, AI-based tools can assist mediators in facilitating communication and developing different types of questions, such as “What do you hope to achieve through mediation today?” Biometric AI technology currently exists to assist mediators in understanding and interpreting human emotions to identify underlying emotions.
Negotiation
Perhaps the most significant application of AI includes the predictive ability based on data collected and what it portends for attorneys using AI to support their negotiation positions. AI-based negotiation tools currently exist to assist the parties, and potentially the mediator, in pursuing favorable negotiated outcomes. Negotiation strategies can be implemented using strategic tools that surpass traditional distributive bargaining exercises, allowing mediators to craft creative solutions to disputes.
Drafting Settlement Agreements
AI can assist in the real-time preparation of settlement agreements and other closing documents, thereby reducing the opportunity for confusion or scrivener errors in a post-mediation drafting exercise.
Mediator Feedback
Perhaps most intriguing is the possibility that someday AI can be used by the mediator to provide critical analysis and feedback, either in real-time or post-mediation. By asking AI reflective questions such as “How did my efforts in today’s mediation compare to other recent mediations?” or “What questions did I pose that elicited the most useful information or informative response during private caucus?” the mediator can improve her skill development.
Ongoing Concerns About the Ethical Considerations of AI in Mediation
As with any new and fast-developing technology, some legitimate questions remain regarding the use and application of AI for mediation. Properly managing this so-called “4th party” in the room will require consideration and potential erection of guardrails to prevent various misguided outcomes. What are a few of the most concerning considerations?
Bias
Perhaps the most troubling of all concerns is that to the extent AI relies on algorithms based on collected data, the predictive outcomes will reflect the potential bias of that underlying data. When the data used to train an AI system reflects past prejudices and inequalities or is not representative of the broader population, an AI’s decisions can be skewed, and the biases may replicated or even amplified, leading to unfair or discriminatory outcomes.
One example of bias offered by a recent conference attendee was that when she (using her female voice) instructed Siri to perform a task, her instructions were ignored. When her husband, using a decidedly deeper, more masculine voice, gave Siri the same instruction, the task was immediately carried out. This is but one small example of potential implicit bias in the manifestation of AI that reflects the larger world bias built into the underlying data.
AI’s machine learning algorithms can also be biased based on their design or the assumptions built into them. This is particularly problematic in mediation, where fairness and impartiality are crucial. AI’s lack of creative and expansive thinking can prevent it from identifying and correcting these biases.
Additionally, AI’s inability to fully understand the complex context of disputes, including cultural, social, and personal factors, can also lead to biased outcomes.
And because AI systems follow predefined rules and patterns, they are less flexible in handling unique or nuanced situations. This rigidity can result in biased outcomes when the AI cannot adapt to the specific needs or circumstances of the parties involved, unlike a human mediator who can think outside the box.
To mitigate these issues, it is essential to ensure close monitoring and updating of AI systems to identify and correct biases and to use diverse and representative datasets for training AI. The judgment of experienced mediators will be needed to complement the use of AI tools to ensure a balanced and fair dispute resolution process.
Ethics
AI can help identify ethical moments in mediation, which offers significant assistance to those not otherwise sensitized to a potential ethical issue. That said, this assistance will need to be premised on standardized ethical rules and ethical considerations. Whose rules will become the standard, and what cultural, political, or legal issues this will surface still need to be addressed. Finally, while many can agree on generalized standards, the nuanced application of those standards may still be beyond AI’s current ability to apply.
Liability
Who is responsible when AI gets it wrong? Many are familiar with the recent example of the New York attorney submitting a legal brief replete with AI-generated misapplied cases. Similarly, a colleague recently used AI to research legal authority to guide his role as Special Master. When he began reading the cases cited to support his intended action, he quickly learned the cases neither supported the proposition asserted or were made up entirely. Commonly referred to as hallucination, AI often seeks to fill in gaps reflecting unavailable knowledge. When confronted, the AI source apologized for getting it wrong.
Legal professionals need to be informed about ethical considerations and apply independent professional acumen to ensure the intelligence part of AI is accurate. The problem of AI occasionally hallucinating responses in lieu of substantively correct answers may be emblematic of early generative AI. Nonetheless, anyone using AI, including mediators, must apply independent professional acumen to ensure the intelligence part of AI is accurate.
Confidentiality
As we collect, store, and access increasingly large amounts of data, how do we ensure the privacy and confidentiality of this data? If not properly protected, the already skeptical, privacy-oriented public will be even less inclined to allow the appropriate collection of data upon which AI depends.
Machines vs. Human Mediators
One cannot engage in a meaningful discussion of AI without considering the ultimate, if not seemingly inevitable, trajectory of AI use in mediation. Many will argue that the choice between AI as “worthless” and AI as “replacing trained mediators altogether” is no longer binary. However, the impact on the human “soft skills” that mediators bring into the room looms large in any debate.
While AI technology can enhance efficiency, provide valuable insights, and increase accessibility, the core human elements that separate mankind from machines – empathy, creativity, kindness, and trust-building, emotional intelligence, and ethical judgment – remain irreplaceable. AI will not replace the ability of an experienced mediator to navigate complex social dynamics, handle the emotional and ethical complexities of mediation, or make nuanced moment-to-moment decisions that clients expect from us.
AI Does Not Replace Formal Mediation Training for Human Mediators
While generative AI tools like ChatGPT can provide mediators with feedback, simulations, and other resources for developing their mediation skills, they cannot replace comprehensive, formal mediation training. (Just ask ChatGPT yourself!) Formal mediation training offers in-depth and thought-provoking lessons from experienced professionals and a structured learning environment that is effective for developing effective mediation skills. AI tools must be viewed as complementary to a trained mediator’s critical thinking skills, not a replacement for those skills.
Moving Forward
As with all change, there will be those who are slow to adapt. Like everything new in mediation, innovation requires a change of mindset. After all, it wasn’t too many years ago that many of us could not envision a role for virtual mediation, and now it has become the default setting for many.
The future will likely see a hybrid approach where AI and human mediators work together, leveraging the strengths of both to improve the mediation process and deliver better outcomes for all parties involved. Mediators who adopt AI as a complementary tool and continue to develop their human-centric skills will be well-positioned to thrive in this evolving landscape.
Hopefully, the days of avatars replacing human mediators aren’t on the horizon; those mediators who embrace the new technology in the room will be one step closer to advancing the superpower of mediation. Those who fail to heed the game-changing potential of artificial intelligence in mediation are in danger of being left behind. Human mediators, with their ability to think strategically and craft creative solutions, will remain essential in addressing the complex, subjective, and emotional aspects of disputes.
