Artificial intelligence is no longer speculative in the legal profession. It drafts, analyzes, predicts. The question for commercial mediators is not whether AI will influence our work. It already does. The question is whether it will replace some of us—and if so, which ones.
Recently, I analyzed the written opinions of nearly two dozen seasoned commercial mediators, most with decades in the field. Their views ranged from confident resistance to cautious adaptation to quiet acceptance of eventual displacement. Taken together, their responses reveal less about the limits of AI and more about the structure of our profession.
But here’s what mediators who feel threatened should understand: this disruption is only truly uncomfortable from the perspective of the status quo. For mediators who have relied primarily on being “the one with the most legal knowledge” or who treat mediation as a formulaic process, AI presents an existential threat. For everyone else—for mediators willing to evolve—this is an unprecedented opportunity. The technology is accessible to everyone. There is no reason any mediator cannot become one of the leaders in advancing the profession to the next level, making themselves genuinely indispensable in the process.
The Fear: Mediation Is Fundamentally Human
Many experienced mediators maintain that mediation cannot be automated. Their argument is substantive, not nostalgic.
Mediation requires trust, rapport, credibility, and human judgment. It involves reading the pause before a corporate representative answers a question about authority. It requires recognizing when anger is a negotiating tactic and when it is masking genuine risk exposure. It demands an ability to absorb competing narratives without flinching and then quietly reshape expectations. And perhaps most critically, it requires the ability to be an emotional center of gravity in the room—someone whose genuine presence makes parties feel something that shifts their willingness to move.
Those skills are not easily reduced to code. Still, if AI can generate sophisticated risk assessments at minimal cost, the middle tier of the mediator market could narrow. Markets do not reward tradition. They reward efficiency.
The Insight: Much of Commercial Mediation Is Structured
We should resist over-romanticizing what mediators do.
A substantial portion of commercial mediation is analytical. We evaluate liability exposure. We quantify damages ranges. We model trial risk. We structure brackets and mediator’s proposals. We help parties make decisions under uncertainty. These processes are structured. Structured processes can be modeled.
Several mediators acknowledged that AI could replicate many dollar-focused bargaining techniques. One described using AI-generated draft awards—one favoring each side—to break impasse. The result was not theoretical. It moved the numbers.
In that respect, it is possible to see a future where at least some cases settle because of AI-mediation—AI does not necessarily need to replicate empathy to affect outcomes in at least some cases. It needs only to influence risk perception. A confident, well-reasoned AI-generated analysis may carry persuasive force even when its underlying reliability is difficult to measure. Framing has always shaped settlements. AI will increasingly shape the framing.
And here is the uncomfortable but useful truth: Parties do not settle just because a mediator is human. They settle because they believe the risk has been clarified and because they believe the deal represents a better alternative when compared to litigation—i.e., it’s a deal that leaves them better off. But there’s another truth running alongside this one: in difficult, high-stakes cases, parties also settle because someone in the room had the emotional gravitas to make them believe settlement was the right thing to do—not just the rational thing, but the human thing. That capacity to make people feel something authentic may be harder to automate than we think.
If a machine can clarify risk convincingly and leave parties feeling that they have done their best to achieve the best possible deal, then that machine will occupy space in this ecosystem. What and how much? As Yogi Berra said, “It’s tough to make predictions, especially about the future.”
The Most Likely Case: Hybridization and Stratification
The most plausible outcome over the next decade is neither displacement nor stasis. It is hybridization, followed by stratification.
In the near term, AI will augment mediators and litigators alike. Lower-value, primarily economic disputes will see partial automation. That is the predictable direction of cost-sensitive markets—cases where the economics do not justify a human mediator.
Where experienced human mediators retain a durable advantage—for now—is in managing hidden drivers: internal authority constraints, insurer-insured tensions, reputational considerations, business relationships, counsel dynamics that do not surface in briefs. These variables are rarely clean. They emerge through disciplined inquiry and conversation. And they often emerge only when someone with genuine presence creates the emotional safety for difficult truths to surface.
Yet even this advantage may be less durable than it appears. AI systems can already detect micro-expressions, analyze speech patterns for evasion or stress, and cross-reference corporate filings, litigation histories, and market conditions in real time. A mediator’s skill in ‘reading the room’ relies on pattern recognition accumulated over years. Machine learning excels at pattern recognition across datasets no human could internalize.
The question is not whether AI can detect that a representative’s bottom-line claim is pretextual, or even whether AI can identify the emotional dynamics in the room. The question is whether AI can make people feel understood, make them believe they’ve been truly heard, and give them the emotional permission to move off entrenched positions. That’s a different kind of intelligence.
The shift to virtual mediation may prove more consequential than most practitioners realize. In-person mediation offered mediators a defensible moat: physical presence, intuitive human connection, the ineffable quality of ‘being in the room.’ Zoom collapsed that distance without replacing it with anything equivalent. It normalized technology-mediated dispute resolution. And in doing so, it may have eroded one of the mediator’s most powerful tools: the ability to be a commanding emotional presence that changes the energy in a room.
That normalization is AI’s beachhead. If parties are already comfortable resolving six-figure disputes over video, the leap to an AI-assisted or AI-conducted session is smaller. We have already trained the market to accept mediation without physical presence. It is difficult to now credibly argue that human presence is indispensable.
The displacement mechanism is already visible. Youth are being taught to use AI as a research assistant, problem-solving tool, and writing collaborator. They will enter the workforce with no memory of pre-AI workflows and no cultural attachment to human-only processes.
When those individuals become decision-makers—GCs, CFOs, claims directors—they will not ask, ‘Should we try AI mediation?’ They will ask, ‘Why are we still paying for humans?’ Nostalgia is not a business case.
But here’s an interesting wrinkle: what happens when both sides are represented by AI? As AI legal assistants become more sophisticated—and as more litigants use AI to navigate disputes without traditional counsel—does the need for human-centered mediation actually increase? Two AI systems negotiating with each other might optimize for mathematical outcomes, but someone still needs to ensure the human beings behind those systems feel the resolution is just. That emotional validation may become more critical, not less, in an AI-dominated legal landscape.
The result is likely to be a sharply stratified profession. Complex, personality-driven, high-stakes disputes will continue to demand experienced human neutrals with established credibility. Specifically, they will demand mediators who can be an emotional center of gravity—people whose presence makes others feel something that shifts the entire dynamic of the negotiation. At the same time, scalable AI-driven platforms will absorb some percentage of matters. Between those poles, competition will tighten.
The Economic Reality
One mediator observed that institutions are sinking trillions into AI and will expect return on investment through job elimination. This is not cynicism; it is industrial logic.
Insurers, corporate legal departments, and ADR providers all have structural incentives to reduce mediation costs. Human mediators require fees, scheduling coordination, travel expenses, and accommodate individual availability. AI systems operate continuously, at marginal cost approaching zero, with no scheduling constraints.
When large institutional users begin mandating AI-assisted or AI-conducted mediation for routine matters the rest of the market will follow. The transition will not be driven by parties choosing AI because it is better. It will be driven by institutions choosing AI because it is cheaper.
So, What’s Left to Do?
For civil litigators and mediators, the lesson is straightforward.
AI is neither a gimmick nor a substitute for judgment. It is a force multiplier for those who understand it and a liability for those who misunderstand it. The advantage will belong to lawyers and mediators who can use AI without outsourcing judgment to it.
Mediation remains the disciplined management of risk and uncertainty in pursuit of durable resolution. AI will increasingly assist in that work. It will compress certain segments of the market and reward adaptability.
But for the foreseeable future, resolving complex civil disputes will still require steadiness, experience, the ability to navigate what does not appear in the data, and the ability to have uncomfortable conversations about reality. And increasingly, it will require something else: the capacity to be a genuine emotional presence—someone who makes people in the room feel something that shifts their calculus.
The mediators who survive won’t just be smart. They’ll be people who can walk into a room and change its emotional temperature. People who can make a hardened litigator pause and reconsider. People who can make a CEO feel genuinely heard and understood. These mediators will likely use AI’s analytical tools to sharpen their preparation and test their assumptions. But their real value will be their emotional gravitas—their ability to be the human center of gravity that makes settlement feel not just rational, but right.
That skill is core to what mediators do. We sit across from people and help them see difficult truths about their cases, their positions, and their alternatives. The mediators who survive this transition will be those who can have equally uncomfortable conversations with themselves about their own value proposition.
The future is unlikely to be human versus machine. It will be human informed by machine—until, in some contexts, it is simply machine.
Authority in that landscape will belong to those who can deliver value AI cannot replicate, in disputes where that value is worth paying for. And the value AI may struggle most to replicate is emotional gravitas—the ability to make people feel something that changes their willingness to settle. For many mediators, that will mean moving upmarket, specializing deeply, or leaving the profession. For litigators, it means understanding that the mediation landscape they navigated for the past two decades is already dissolving.
The mediators who believe they are indispensable should ask themselves: Would my clients pay 10x more for me than for a comparable AI system? If the answer is anything other than an immediate yes, the question is not whether they will be affected. The question is when.
But here is the opportunity: the mediators who answer that question honestly—and then do the work to become genuinely worth 10x—will not only survive this transition. They will lead it. They will define what mediation becomes in the age of AI. They will master the technology and deploy it strategically. But their real edge will be offering something no algorithm can replicate: genuine emotional presence, the ability to make people feel understood and safe enough to move, and seasoned judgment about when human gravitas matters more than machine precision.
That is not a niche market. That is the future of the profession.
Even so luminary a firm as O'Melveny has been smacked down by the courts (here, the Ninth Circuit) when trying to enforce employee arbitration agreements. California lawyers would therefore be...
By Victoria PynchonCheck out reporter Kevin Fogarty's article How to Turn a Job Offer into a Raise over at The Ladders, an Executive Search company for the six-figure set. Kevin recently interviewed...
By Victoria PynchonSid Lezak discusses the concerns he has had regarding the field. These include overly extravagant claims made about what the field brings to clientele, over-promising those who are coming into...
By Sid Lezak
