For an established dispute-resolution organization that is turning 100 next year, the American Arbitration Association seems to be doing everything but acting its age. Long among the world’s leading providers of human arbitrators and mediators for a range of disputes, the AAA is now preparing to launch its first AI-powered arbitrator in November.

The AI arbitrator will initially handle documents-only construction cases, a high-volume area where the organization sees efficiency and speed as particularly valuable. The system is designed to automatically evaluate case merits, generate recommendations, and prepare draft awards — to be reviewed by human arbitrators before they are issued.

In an interview for my LawNext podcast that will air next week, Bridget Mary McCormack, the AAA’s president and CEO, projected that the AI arbitrator could reduce the cost to parties of construction arbitration by 30 to 50 percent and the time required for a case by 25 to 30 percent at launch, with those metrics improving as the technology advances.

Eventually, the AAA plans to expand the AI arbitrator into other types of disputes. Diana Didia, executive vice president and chief technology and innovation officer, in that same LawNext interview, said that the next area of development will be insurance cases, and specifically payer-provider disputes, where there is a high volume of cases, usually involving claims of lower dollar amounts.

Trained On Actual Cases

The AI arbitrator was developed through an extensive training process using more than 1,500 actual construction awards from AAA-ICDR’s case repository. The system was specifically designed around legal reasoning as its foundation, Didia said, with human arbitrators providing input throughout the development process.

The training dataset was chosen strategically, Didia said. Construction cases typically include reasoned awards that allow the AI system to map the decision-making chain of thought that human arbitrators use when analyzing evidence and reaching conclusions.

Although the system — which was developed in collaboration with QuantumBlack, AI by McKinsey — has not yet been tested in actual cases, the AAA has tested it across over 1,000 simulated cases using actual completed disputes from AAA’s data repository, with arbitrators, lawyers, and law school students playing the roles of the parties and neutrals in the disputes.

Human-in-the-Loop Framework

A key component of the system is its “human-in-the-loop” framework. After the AI generates its draft decision, human arbitrators review the results, with full access to the case materials, and can revise the AI-generated decisions before they are finalized.

When parties submit their materials, the AI system deconstructs their submissions, identifying claims, evidence and legal frameworks. Crucially, this analysis is presented back to the parties for validation.

“The parties get to say, yes, that’s right, or no, this part’s wrong,” McCormack said, “and they get to move it to a place where they are satisfied that it fully understands what they think their case is.”

This validation step is a fundamental game changer compared to traditional dispute resolution, McCormack said, where parties often feel that decision-makers did not understand or address their most important arguments. The AI arbitrator’s transparent breakdown ensures parties know they have been heard and understood before any decision is rendered.

Once all submissions are complete, a human arbitrator from AAA’s permanent panel is appointed through a traditional round-robin system, maintaining the same disclosure and conflict-checking procedures as would apply in any AAA case.

‘A Muscular Co-pilot’

The appointed arbitrator gets access to what Didia described as “a very muscular co-pilot” interface, featuring organized case summaries, timeline views, claims analysis, and, crucially, a complete draft award.

The system gives arbitrators access to all relevant evidence in an organized format, allowing them to click through to source documents while reviewing the AI’s reasoning.

Once the human arbitrator reviews the draft award and supporting analysis, the arbitrator can make any adjustments and edits. These modifications feed back into the AI system for continuous improvement. The final, issued award carries the arbitrator’s name and certification — they are issuing the decision, not merely reviewing an AI output, the AAA emphasizes.

Arbitrators participating in the testing process reported reviewing case materials 30 to 50 percent faster than they would normally, while still maintaining confidence in the outcomes.

“The AI is issuing an award, but the human is validating it and the human is signing,” Didia said.

Potential to Enhance Access

The AI arbitrator’s benefits to litigants could be substantial. McCormack estimates that the process could cut the cost of construction cases by 30 to 50 percent and the time by 25 to 30 percent.

Beyond the immediate benefits to the parties, the AI arbitrator, by reducing costs and complexity,  could open dispute resolution to parties who currently cannot afford traditional processes.

“By bringing the cost down and the time down and also making the process just simpler for users, it’s going to mean that if right now we resolve half a million disputes a year, we can resolve 10x that,” McCormack said. “That’s amazing, right? And it’s just going to open up completely new frontiers for dispute resolution.”

Impact on Arbitrators

Does the advent of AI arbitrators mark the beginning of the end for human arbitrators?

“I don’t think so at all,” McCormack told me, but they will have to change their business models.

“I do think that you want to be able to be one of those arbitrators who’s able to accommodate an AI-native process because you’re going to continue to have a big career,” she said.

Because AI is going to open the arbitration process to more people, she emphasized, that means more opportunities for arbitrators.

“It’s not going to put you out of business. Quite the opposite. It’s going to allow significantly more disputes to come our way.

“But you do have to come along for the ride. I do think that if you’re not along for the ride, then you might want to think about what the next phase of life looks like.”

A Century of Innovation

The AAA’s launch of the AI arbitrator represents the culmination of a broader generative AI and innovation initiative it has been pursuing for the past several years, and which has already produced several tools, including chatbots for rules and customer support, AI-enhanced panelist search capabilities, document summary and Q&A functions, and various case management enhancements.

It also comes as the AAA approaches its centennial anniversary. In these latest innovations, McCormack draws parallels to the organization’s founding, noting that arbitration itself was considered an innovation 100 years ago, designed to provide broad access to dispute resolution for all parties, not just large businesses.

“We think this technology allows us to deliver on that in a modern world,” she said. “Arbitration is a really important process for lots of users, but the world has gotten significantly more complicated and we need more options, and we’re ready to deliver.”

[Disclosure: I am an arbitrator listed on the AAA’s labor relations roster. I receive case appointments through the AAA, but any compensation I receive is paid by the parties to the dispute, not by the AAA.]

Photo of Bob Ambrogi Bob Ambrogi

Bob is a lawyer, veteran legal journalist, and award-winning blogger and podcaster. In 2011, he was named to the inaugural Fastcase 50, honoring “the law’s smartest, most courageous innovators, techies, visionaries and leaders.” Earlier in his career, he was editor-in-chief of several legal publications, including The National Law Journal, and editorial director of ALM’s Litigation Services Division.