Designing the Future of ODR: The ‘Centaur System’ and Role of Humans in the AI Era

Share:

This session from the 2026 Global ODR Forum at Harvard addresses issues of “Designing the Future of ODR: The Role of Humans in an AI-Driven Era.” This session features Renee Jackson, CEO of Dyspute.ai; John Lande, Professor Emeritus at the University of Missouri School of Law; and Morgan Tregenza, CEO at Levelheads Mediation.

The panel explores the “centaur system” concept, where humans and AI collaborate along a spectrum of control to enhance the mediation process. The speakers showcase specific innovations, such as the RPS Coach for negotiation training and Dyspute.AI, an automated platform designed to improve access for high-volume, low-value cases.

Key considerations include the importance of human-in-the-loop models, the necessity of ethical guardrails, and the potential for AI to reduce participant anxiety through non-judgmental interaction. Ultimately, the panel emphasizes that, while technology can democratize access to justice and provide valuable data, maintaining the “magic” of human empathy remains a vital challenge.

The Centaur at the Table: Why the Future of Conflict Isn’t Just Human—It’s Hybrid

Blog Post by Jim Melamed

1. Introduction: The Broken Gavel and the Digital Olive Branch

Our current legal system is failing. It’s a harsh reality, but for the average person, the traditional judicial mechanism is less a hall of justice and more a labyrinth of frustration. Mo Trenza, a pioneer in the field, often shares her origin story of mediating cases in the cramped, dimly lit janitor’s rooms of Colorado courthouses. Even back then, the realization was clear: a judge’s verdict is often practically useless. It’s a blunt instrument applied to a delicate problem.

This is why we talk about the “magic of mediation.” When you bring people together to tell their stories, you create durable, reliable outcomes that a court order simply cannot replicate. But we are facing a massive scalability crisis. To solve it, we must embrace technology not merely as a tool for Zoom calls, but as a “fourth party” at the table. The digital olive branch isn’t just a medium; it’s becoming an active participant in the room.

2. The “Centaur” Strategy: Moving Beyond the “AI vs. Human” Binary

We need to stop viewing the future of conflict resolution as a zero-sum war between human empathy and machine logic. Instead, we should look to what John Lande calls the “Centaur System.” Much like the mythological creature—human head, horse’s body—this model suggests that the hybrid of human intuition and AI processing power is exponentially stronger than either acting alone.

The Centaur operates on a “0 to 5” continuum (a scale conceptualized by Chris Draper). At “0,” there is no AI assistance; at “5,” the AI has complete control. Lande’s philosophy, embodied in his RPS (Real Practice System) Coach, leans toward a “human-in-control” model. This isn’t just having a human “in the loop” to rubber-stamp AI decisions; it’s a system where the AI acts as a sophisticated mentor, respecting party autonomy and supporting professional practice while the human maintains the reins. As Lande puts it:

“The idea is that the combination is stronger than either the human or the horse
… you have a human and an AI entity in the tail.”

3. The Bias Paradox: Why We Sometimes Trust Bots More Than People

Innovation often yields counter-intuitive results. Renee Jackson, founder of Dyspute.AI, conducted a fascinating experiment at the University of New Hampshire, pitting her AI mediator, “Adri,” against human mediators in identical fact patterns. The assumption was that the humans would win on “soul.” The reality? Students rated the AI as more neutral, fair, and unbiased.

This reveals a profound psychological barrier: “judgment anxiety.” We found that people often feel safer “speaking their truth” to a machine. A bot doesn’t have a “tell.” It doesn’t have a bad day or a capacity for condescension. Participants felt they could ask “dumb” questions or express raw emotion without the fear of being looked down upon by a professional in a suit. Jackson’s experiment proved that for many, the machine is the ultimate safe space:

“Students thought that the humans were more biased than our AI mediator… they liked negotiating or mediating with AI because they could say whatever they wanted. They knew they weren’t going to be judged.”

4. The “Lossless” Filter: Solving the Rage-Typing Problem

If judgment anxiety is the psychological barrier, “rage-typing” is the procedural one. In high-conflict disputes, the path to resolution is often buried under a mountain of inflammatory language. This is where the Centaur strategy provides a technical solution to a human problem through “lossless factual transformation.”

On the Adri platform, the AI acts as a professionalizing buffer. A party can vent, scream, and type insults into the interface. The system then filters this delivery through rigorous guardrails trained on the ICODR.org and ABA ethics standards. The “lossless” element is crucial: the AI strips away the rage, but leaves the factual demand and the core sentiment intact. It ensures the other party hears, “I am frustrated because I haven’t been paid,” rather than the vitriol that originally accompanied it. By filtering the delivery while preserving the substance, we keep the path to resolution open when it would otherwise be incinerated by emotion.

5. The Ivory Tower Resistance: The Danger of Banning AI in Law Schools

Despite this potential, we are seeing a dangerous retreat into the “Ivory Tower.” Institutions like the University of California Berkeley have moved to ban or heavily restrict AI in legal education. This “tenure-protected” resistance is not just short-sighted; it’s a disservice to the profession.

We are already seeing the consequences of this literacy gap. Lawyers in the real world are being sanctioned and even disqualified from cases for failing to manage AI “hallucinations.” If we don’t teach students how to responsibly manage these systems, they will enter the market as liabilities. Resistance to technology doesn’t protect the law; it leaves future clients vulnerable to the technical dysfunctions of an untrained bar. We must move from banning the tool to mastering the machine.

6. The 2:00 AM Mediator: Democratizing Access through Availability

The most transformative aspect of AI-driven ODR (Online Dispute Resolution) is the demolition of barriers to access. For the “high-volume, low-dollar” disputes that clog our courts—evictions, small business tiffs, minor consumer claims—the cost of a human mediator is an impossible wall.

The AI is the “mediator in your pocket,” available at 2:00 AM when a tenant’s anxiety is highest, or while someone is waiting in line for groceries. It doesn’t require a scheduled session in a downtown office. It meets people where they are, on their own timeline. This democratization is how we finally fix the “broken gavel”—by providing a resolution mechanism that is as accessible as a smartphone app.

7. Conclusion: Can We Code the “Magic”?

We are at a crossroads in the history of dispute resolution. The tension between the efficiency of the machine and the “magic” of human rapport is real. But as we’ve seen, the Centaur model allows us to have both. We don’t have to choose between a cold bot and an unreachable human expert.

The future of conflict resolution lies in our willingness to let the machine handle the “rage” so the humans can focus on the “rapport.” We can use AI to provide the neutrality, the 2:00 AM access, and the factual filters, freeing us to focus on the deep, transformative work of human connection.

Final Thought: Are we brave enough to trade our attachment to the “magic” of human empathy for a system that is demonstrably more fair, accessible, and objective?