The promise, the limits, and what it really means to help people find common ground.
Mediation is one of the oldest human arts. Long before courts and contracts, communities relied on a trusted person to sit between two quarreling parties, hear them out, and guide them toward something both could live with. Today, artificial intelligence can draft contracts, diagnose diseases, and write symphonies. So, the question is worth asking seriously: can it do this too?
The answer is nuanced. Where you land on it says a lot about what you think mediation actually is.
What Mediators Actually Do
Before we can ask whether AI can replace a mediator, we need to be honest about what the job involves.
At its surface level, a mediator facilitates communication. They set ground rules, manage speaking time, summarize positions, and propose frameworks for resolution. This part looks process-driven and rule-governed, exactly the kind of task AI handles well.
But beneath the surface, something harder is happening. A skilled mediator reads the room. They notice that one party goes quiet when money is mentioned, or that the real grievance isn’t about the contract clause but about feeling disrespected. They choose when to push, when to let silence do the work, and when to say something that reframes the entire conflict. They hold emotional weight without absorbing it. They build trust simultaneously with two people who distrust each other.
That is a very different skillset, and it is much harder to automate.
Where AI Is Already Showing Up
AI is not waiting for permission. It is already inside dispute resolution in several meaningful ways.
Online Dispute Resolution (ODR) platforms now handle millions of low-stakes cases. eBay alone uses algorithmic resolution for over 60 million disputes per year, more than the entire U.S. civil court system combined.[1] These systems route claims, match offers, and close cases without a human mediator ever getting involved. For simple, transactional conflicts with clear facts and quantifiable outcomes, they work remarkably well.
Document analysis and preparation is another frontier. AI can review contracts, flag ambiguous clauses, identify points of likely contention, and brief a mediator before a session begins. This doesn’t replace the mediator. It makes them sharper.[2]
AI-assisted negotiation tools can model outcomes, suggest settlement ranges based on comparable cases, and give parties a data-grounded sense of their alternatives. This is particularly valuable in commercial disputes where information asymmetry is a core problem.
In each of these cases, AI is doing what it does best: processing information quickly, removing friction from low-complexity interactions, and helping humans make better-informed decisions.
Where AI Falls Short
Mediation is not just information processing. Consider a divorce mediation where two people are grieving the end of a marriage while simultaneously negotiating custody of their children. The most important moment might be a pause. A hand gesture toward the tissue box. The mediator saying, quietly, “It sounds like what you really want is for him to acknowledge how hard this has been.” That moment requires presence, intuition, and human recognition of human pain. Experts at Harvard’s Program on Negotiation note that generative AI remains ill-equipped to help parties cope with the strong emotions that typically arise in mediation.[3]
Trust is earned through presence. People in conflict are often hypervigilant. They need to believe that the person across the table from them is genuinely neutral, genuinely listening, and genuinely invested in a fair outcome rather than running an algorithm. Research published in the Conflict Resolution Quarterly found that when cases involved intense human emotions, participants had significantly lower acceptance of AI involvement and preferred human mediators.[4] The perception of humanity matters, not just its simulation.
Power imbalances are felt, not computed. In many mediations involving workplace harassment, family disputes, or negotiations between a large corporation and an individual, one party is vulnerable. A human mediator can sense when someone is being coerced into agreement, when a “settlement” is really capitulation, and when the process itself needs to stop. This is an ethical judgment, not a data point.[5]
Cultural and contextual fluency is irreducible. The meaning of a gesture, a tone, a silence: these are embedded in culture, history, and relationship. A mediator working in a specific community over many years develops a kind of contextual intelligence that no model trained on generic data can replicate.
The Better Question
Perhaps “can AI replace a mediator?” is the wrong question. A sharper one: where should AI take the lead, and where should humans?
For high-volume, low-complexity disputes such as consumer claims, minor contract disagreements, and straightforward neighbour conflicts, AI-driven ODR may be not just acceptable but preferable. It is faster, cheaper, more consistent, and removes the human biases that sometimes infect traditional mediation.
For emotionally complex, high-stakes, or power-imbalanced disputes such as family law, workplace conflicts, community trauma, and cross-cultural negotiations, human mediators remain essential. AI can be a powerful tool in their hands but should not be the one sitting at the table. Most current experts and platforms position AI as an assistant, not a sole mediator.[6]
The risk is not that AI will take over human mediation overnight. The risk is subtler: that the efficiency and scalability of AI-driven resolution will quietly push out human mediators in cases where they were needed, because the AI option is cheaper and the harm is invisible.
The India Dimension: A Crisis That Cannot Wait
Nowhere is the case for AI-assisted dispute resolution more urgent than in India. As of June 2026, over 56 million cases are pending across the country’s courts, with more than 180,000 of them sitting unresolved for over 30 years. [7]At the current pace of disposal, experts estimate it would take several hundred years to clear the backlog entirely.[8] The average judge is managing over 2,200 cases simultaneously. This is not a system under strain. It is a system in collapse.
For millions of ordinary Indians, this means justice deferred is justice denied. An acid attack survivor waits a decade for a verdict. A land dispute outlives the people who filed it. A small business owner cannot recover dues because a commercial case drags through three courts over fifteen years. The human cost is immeasurable, and it falls hardest on those who can least afford lawyers or long waits.
The legislative response has arrived. The Mediation Act, 2023, India’s first standalone legislation on mediation, explicitly recognizes online mediation as a valid and enforceable process.[9] It mandates pre-litigation mediation in many civil matters, establishes the Mediation Council of India, and makes mediated settlement agreements enforceable as court judgments once registered. This is a foundational shift. It means AI-assisted ODR is no longer a fringe experiment. It has legal standing.
NITI Aayog lit the fuse. In 2021, NITI Aayog released its ODR Policy Plan for India, recommending that online dispute resolution be integrated into e-commerce, banking, and consumer grievance sectors through public-private partnerships.[10] The Reserve Bank of India followed with a mandate requiring ODR for failed digital payment disputes. With India processing billions of digital transactions annually, this alone represents a staggering volume of potential cases suited to AI-assisted resolution.
Private platforms are already building the infrastructure. Indian ODR platforms including Track Second, SAMA, Presolv360, CADRE ODR, and TERES are operationalizing AI-assisted mediation and arbitration across commercial and consumer disputes.[11] These are not pilots or proofs of concept. They are live systems handling real cases, many of which would have otherwise joined the court backlog.
India also has something the global conversation often overlooks: its own mediation tradition. Panchayats, community elders, and informal dispute settlement have existed here for millennia. The idea that a neutral third party can help two sides find common ground is deeply embedded in Indian culture. AI does not threaten this tradition. It could extend it, making access to structured, fair, and timely resolution available to communities far from urban legal infrastructure, in regional languages, at negligible cost.
But the challenges are real. India’s digital divide means that ODR platforms primarily reach the urban, the educated, and the digitally connected. Rural litigants, who form the majority of the backlog, often lack reliable internet access, smartphones, or the digital literacy to navigate an online platform. Ensuring that AI-assisted mediation reaches Bharat, and not just India, will require deliberate policy investment alongside the technology.
There is also the matter of algorithmic transparency. Legal scholars have noted that India currently lacks a coherent regulatory framework governing AI in dispute resolution, raising concerns around bias, data privacy under the Digital Personal Data Protection Act 2023, and accountability when an AI-assisted process produces an unjust outcome.[12]
The opportunity is historic. With over 56 million pending cases, a freshly enacted Mediation Act, a growing ODR ecosystem, and a government committed to digital justice, India stands at a genuine inflection point. AI will not clear the backlog alone. But deployed wisely, with human mediators leading the complex cases and technology handling the high-volume low-stakes ones, it could deliver something India’s judiciary has struggled to offer for decades: timely, affordable, accessible justice.
What This Means for the Profession
Mediators who treat AI as a threat are missing an opportunity. Mediators who treat it as a full replacement are making a category error.
The most effective practitioners of the next decade will be those who harness AI for what it does well: research, pattern recognition, process management, documentation. They will pair that with the uniquely human capacities no model can replicate, including presence, ethical judgment, emotional attunement, and the hard-won wisdom of having sat with human conflict long enough to understand its grammar.
Empathy is not a feature. It is not a function call. It is the irreducible core of what makes mediation work, and for now, it remains stubbornly, beautifully human.
The question was never whether machines could negotiate. It was whether they could understand why people fight in the first place. Authored by Kunal Peelwan, Advocate & Mediator
[1] Rule, Colin. “Online Dispute Resolution Moves From E-Commerce to the Courts.” The Pew Charitable Trusts, June 4, 2019. https://www.pew.org/en/research-and-analysis/articles/2019/06/04/online-dispute-resolution-moves-from-e-commerce-to-the-courts. Rule served as Director of ODR for eBay and PayPal from 2003 to 2011.
[2] Mediate.com. “AI Empathy in Mediation: When Algorithms Show Compassion.” January 14, 2026. https://mediate.com/ai-empathy-in-mediation-when-algorithms-show-compassion
[3] Poole, Christopher K. (CEO, JAMS). Quoted in: Program on Negotiation, Harvard Law School. “AI Mediation: Using AI to Help Mediate Disputes.” May 6, 2026. https://www.pon.harvard.edu/daily/mediation/ai-mediation-using-ai-to-help-mediate-disputes
[4] Choi, Y. et al. “Using AI in My Disputes? Clients’ Perception and Acceptance of Using AI in Mediation.” Conflict Resolution Quarterly 43, no. 2 (2025): 223–238. https://doi.org/10.1002/crq.21483
[5] Tabor, Francesca. “The AI Mediator: Can AI Help with Dispute Resolution and Peace Keeping?” June 3, 2025. https://www.francescatabor.com/articles/2025/6/3/the-ai-mediator-can-ai-help-with-dispute-resolution-and-peace-keeping
[6] Pollack Peacebuilding Systems. “AI Driven Mediation: Best Practices and Future.” April 19, 2026. https://pollackpeacebuilding.com/blog/ai-driven-mediation
[7] Wikipedia. “Pendency of Court Cases in India.” Updated June 2026. https://en.wikipedia.org/wiki/Pendency_of_court_cases_in_India. Data sourced from the National Judicial Data Grid.
[8] The Christian Science Monitor. “54 Million Cases, and Counting: Why India’s Court Backlog Keeps Growing.” March 18, 2026. https://www.csmonitor.com/World/Asia-South-Central/2026/0318/India-court-backlog-delayed-justice. Experts cited include Shruthi Naik, Head of the Justice, Access and Lowering Delays in India initiative at the Vidhi Centre for Legal Policy.
[9] Ministry of Law and Justice, Government of India. The Mediation Act, 2023 (No. 32 of 2023), receiving Presidential assent on September 14, 2023. https://legalaffairs.gov.in/sites/default/files/MediationAct2023.pdf
[10] NITI Aayog. “Designing the Future of Dispute Resolution: The ODR Policy Plan for India.” 2021. https://www.niti.gov.in/sites/default/files/2023-03/Designing-The-Future-of-Dispute-Resolution-The-ODR-Policy-Plan-for-India.pdf. Also cited in: Shardul Amarchand Mangaldas. “From Litigation to Automation: The Legal Future of ODR in India.” https://www.amsshardul.com/insight/from-litigation-to-automation-the-legal-future-of-odr-in-india
[11] IRCCL. “Beyond the Gavel: Trends, Hurdles, and Future in Online Dispute Resolution.” March 24, 2026. https://www.irccl.in/post/beyond-the-gavel-trends-hurdles-and-future-in-online-dispute-resolution
[12] EPRA Journals. “A New Era of Mediation.” November 2025. https://eprajournals.com/pdf/fm/jpanel/upload/2025/November/202511-01-024736.The paper calls for amendments to the Mediation Act 2023 to include an AI governance framework, and for the Digital Personal Data Protection Act 2023 to cover AI data processing in legal and mediation contexts.