In Episode 19 of the AAAi Podcast, hosts Bridget McCormack and Zach Abramowitz focus on the topic that now defines serious AI deployment in legal services: governance. The American Arbitration Association’s AI Arbitrator has drawn national media attention, but the real story is not novelty — it’s structure. As AI tools rapidly mature, the technology is delivering real value, increasing efficiency, and creating measurable operational gains for organizations willing to use it. The differentiator is no longer whether the technology works, but whether institutions are building transparent, accountable systems that people can trust.
That same theme will continue at Legalweek 2026 in New York, where Bridget will moderate an in-person panel titled, “Trust Is the Product: How Proven AI Governance and Legal Oversight Is Driving Successful Adoption.” The session features Anna Gressel of Freshfields, Henry Hagen of Moderna, and Galia Amram of OpenAI. Taking place Tuesday, March 10, 2026, from 11:30 AM – 12:30 PM, the panel will explore how governance frameworks — not hype — are shaping durable AI adoption inside legal departments and institutions.
Key Takeaways
- The AI Arbitrator Has Entered the Broader AI Conversation
The AI Arbitrator is no longer just a legal-tech development — it has entered the mainstream AI discussion. National outlets, including the Wall Street Journal and the Verge, have begun covering the AAA’s work because it sits at the intersection of institutional trust, public accountability, and applied AI. What distinguishes this effort is not simply automation, but the commitment to systems that show their reasoning and operate within established legal standards.
This broader attention reflects something important: AI in adjudication is no longer hypothetical. It is being built by established institutions with long-standing credibility. When AI systems are transparent, auditable, and deployed selectively where appropriate, they move from experiment to infrastructure.
- Scaling Agency Through Agentic Systems
At its core, the AAA’s AI Arbitrator is about expanding access to dispute resolution. The AI Arbitrator is an agentic system trained on AAA expertise and prior awards, designed to provide resolution options for disputes that historically may have gone unresolved because they were too small, too costly, or too inefficient to pursue.
This approach is not a departure from AAA’s founding principles — it is an extension of them. From the beginning, the organization focused on giving individuals and businesses greater agency over how their disputes are resolved. AI, when deployed responsibly, allows that opportunity to scale. It creates additional pathways for resolution while maintaining institutional guardrails.
Importantly, this does not mean AI is appropriate for every dispute. Matters involving public rights or government accountability belong in public courts. But for many private disputes, a transparent and carefully governed AI system can offer faster, more accessible options without compromising fairness.
- Governance IS the Product
AI adoption is no longer a theoretical question. Organizations are already using these tools, but what separates responsible deployment from ad hoc experimentation is governance.
Governance in this context means operational discipline: clear policies, documented oversight, transparent outputs, audit trails, and regular evaluation. It requires answering practical questions about how systems are monitored, how results are validated, and how accountability is maintained.
In 2026, competitive advantage in AI will not come from simply “using AI.” It will come from proving that its use is structured, supervised, and measurable. Institutions that treat governance as central — not as an afterthought — are the ones building durable trust.
- The Walmart Model of AI Strategy
Not every organization needs to build its own large language model to benefit from AI. Bridget highlighted Walmart as an example of intelligent adoption: applying existing AI tools strategically to improve user experience and operational efficiency rather than attempting to own the entire technology stack.
The lesson for legal organizations is similar. Competitive strength does not require building foundational models from scratch. It requires understanding how to apply powerful tools thoughtfully, so that they enhance — rather than degrade — the experience of the people they serve. Strategic integration, not technical ownership, is what drives value.
Final Thought
The AAA’s leadership in AI stems from aligning technology directly with its founding mission: expanding access to fair and efficient dispute resolution. When innovation reinforces institutional purpose, it becomes an accelerator rather than a distraction.
We are in a transitional moment. The organizations that succeed will be those that move beyond experimentation and commit to governance frameworks that convert AI innovation into measurable outcomes. In this environment, trust is not a byproduct of technology — it is the product.
Transcript
Zach Abramowitz: So Bridget, very excited to be back here for season three of the AAAi Podcast. It's really incredible to think that we've been on this journey now for as long as we have.It feels to me like AI just got here. In one sense it feels like it's always been here because I can't imagine working without it. But then I also think about it as a recent arrival. I've been talking more recently and mentioning to folks about November 2022 when ChatGPT launched. We're in 2026 now.
Very excited to be back on this journey and I think there's going to be a lot of folks showing up to Legal Week in New York this year to get a better sense of how the profession is being impacted by AI. Obviously something we've spoken quite a bit about here on this podcast, and you're going to be on a panel and I wanted to flag it for all of our audience, first of all, is an opportunity to meet you in person, but also to hear what I think is going to be a very interesting conversation.
You are on a panel entitled Trust is the Product, How Proven AI Governance and Legal Oversight is Driving Successful Adoption. It is at Legal Week, on Tuesday March 10 11:30am. The other folks on that panel with you are Anna Gressel, who's a partner in the Global Co-Head of AI at Freshfields. They've got a very interesting program. They're one of the firms to really run with Google before others have done it. Very interested in hearing her insights, as well as Henry Hagen, Associate Counsel at Moderna. Moderna obviously also one of the first OpenAI partners very early on and their legal department was one of the first groups within that company to adopt AI at scale. So again, very interested in the insights there. Galia Amram, Associate General Counsel of Open AI, who I'm sure folks will be very interested to hear from as well.
Zach Abramowitz: So you are probably in the early planning stages of this panel, but give us a little bit of a sense of what people can come and expect to hear and some of the conversations you think that are going to be taking place that day.
Bridget McCormack: Yes. It is great to be back Zach and I feel exactly like you do, in some ways, 2022 feels like forever ago and in other ways it feels like yesterday. How much has everything changed since then, if you really stop and think about your day-to-day life today versus October of 2022. It's kind of stunning, right? Head spinning and in some ways, I'm really excited about this panel at Legal Week mostly because of those spectacular panelists. I'm moderating this conversation and we're focusing on governance because I think that's where the conversation is now, in legal AI. Everybody's using AI some are using it pursuant to a strategic plan, a sophisticated approach within their legal team, legal departments. Others are still using it at home or on their phone and sneaking it in to work. That's still happening in some places.
Bridget McCormack: Everybody is figuring out governance as we go.The AAA is no different in that way. In a way we were building our governance process as we were building our AI point solutions. And then some of these larger services that we're now offering that are AI native the technology is already bringing value and can bring significant value to anybody who wants to use it. The people are along for the ride and they have to trust it. And so that's all about your governance, right?
How transparent are you with what you're building? How do you show your users, your stakeholders, what you're doing, how you're doing it? What does your audit trail look like? How often are you releasing those results? For many firms and other organizations, that's still pretty nascent. Folks are still kind of figuring that out and this is a great panel for that because these are very sophisticated thinkers about the governance around AI. I had this conversation with Chris Mims at the Wall Street Journal a couple months ago, he was talking about it as the harness, which I think is how the technologists talk about governance.
Bridget McCormack: We lawyers call it governance, the engineers call it, the appropriate harness that you put around your agents or whatever it is you're building and whatever your name is for it. In some ways, it's going to be the most important part of what you're building in the next few years because that's the only way the people you are bringing along for the ride, they need to trust it. Maybe that all changes in 10 years when we get through this liminal period. But getting through the liminal period requires a real commitment to your governance process around what you're building with AI.
So that's why this is going to be a fun conversation. It's also a topic we want to put out a lot more thought leadership about. We're going to put a survey in the field of lots of C-suite leaders about their AI governance so that we can start talking about where folks are, and bring people along. because I do think it's an important leg of the stool in getting us to widespread adoption.
Zach Abramowitz: Yes. A couple of points on governance and where that's prioritized in legal departments today. So I've been reading the 2026 CLO survey, which I read very, very closely every year. I tend to look for deviations. because, for years when you would read the survey, mostly everything felt like it was staying the same. You might see legal operations as a priority making more inroads. A survey that suggested that many legal departments are pursuing new technology. But there was something very compelling in this year's surveys. Number one, technology implementation is a strategic priority. Last year, we already noted that it was up from 6% to 14%. Answering that it was a strategic priority, and it was the second biggest answer, technology implementation is a strategic priority.
In this year's survey, 38% suggested that it was the most important strategic priority, the number is already up. But what was interesting right ahead of that was regulatory compliance. And I think that this is absolutely the case if you go into legal departments today, yes, there is definitely an emphasis on AI adoption. But I think that it actually outpaces what's going on at law firms, which is to me a huge surprise. I would not have thought it would go that way, but the other thing that I think is just super telling is and I've seen this at multiple, very large legal departments where I would speak to them in 2024, and I would speak to the lawyer put in charge of infusing AI into the legal department. That was where their heads were at. 2025 when I would speak to them. It's no, I've stuck someone else on that. Right now, our legal department is busy making sure that the compliance and guardrails and governance around the agents that are being rolled out through the rest of the company, compliance and that is this entire new challenge that legal departments have on their hands.
So of course we're adopting AI. Because the sophistication of our roles, number one, you know, has increased dramatically because we're dealing with entirely new territory here that we've got to be on top of. So, of course we're adopting AI tools, we need all the help that we can get. More than that, we are focused on governance. So that seems to be borne out in both an industry survey, and anecdotal data. So I think obviously, that's going to be a critical topic moving forward.
Bridget McCormack: Yes, I think it's going to be a pretty interesting conversation and I highly recommend if you're at Legal Week, you stop by. The panelists are fantastic and it's something we turned to in 2025, but more and more are turning to in 2026. And so I'm excited to start this conversation and then keep at it throughout the year. It's going to be a good one.
Zach Abramowitz: Yes, I think towards the end of the year and the beginning of this year, obviously since I pay a lot of attention to what's going on at the American Arbitration Association. I couldn't help but notice that there's been some really significant news coverage of The American Arbitration Association over the last few years since you've begun investing in AI and really building in public both through this podcast, but other major thought leadership, you're speaking somewhere else every other week in some other different part of the world.
Zach Abramowitz: You've always gotten coverage with the legal industry press and I think very positive coverage of the AI efforts, Bob Ambrogi and Law.com, Law360, Artificial Lawyer, you've been a really known commodity, and I think that's just generally been true. If you think about the AAA, brand awareness in legal is off the charts. If you go outside of legal, many people haven't necessarily heard of the AAA, or realize that the AAA is a nonprofit, but also a real sustainable business. All of the coverage recently has been in the mainstream press, so twice in the Wall Street Journal in the span of a couple of weeks. Also, a major writeup in the Verge called All Rise for Judge, GPT. Then a notable mention in Business Insider that I'll talk about., I wanted to go through this and get a sense of the different coverage and where it's coming from. But really beyond that, I want to know about your impressions of why this is an area where all of a sudden a company that may have been like an unknown in broader business circles, is now squarely in the news. Why is it that people are following the AAA for its AI efforts right now? Why is that making such waves?
Bridget McCormack: I think it's because even though outside of legal, I agree with you, the American Arbitration Association, isn't a household name. In fact. If you say the AAA, they ask you about the maps, they think it's the triptychs. But what's not outside the mainstream is the cost of disputes and resolving them. And it's something that lots of folks have experience with and kind of dread and try to not think about.Small disputes, medi disputes, large disputes have a track record of taking businesses down, really causing stress and anxiety and our current models for resolving them.
Pre AI, the public dispute resolution system, also known as the courts and the private dispute resolution systems, which is, you know, alternative dispute resolution, arbitration, mediation. You can resolve a dispute any way you want. If you agree to it in a contract. All of those were effective as far as they could go, but an awful lot of disputes were simply unresolved or have been unresolved. So there's kind of an enormous opportunity to make a real difference at the level of the individual, at the level of a community in an enterprise by meeting what is an enormous need. There's all kinds of advantages to be made with AI and legal across the business of law, across the practice of law. But you've heard me say this many times. I think perhaps the most exciting place where AI can make an enormous difference is by scaling dispute resolution and making it accessible to anybody who has a dispute and wants to resolve it. You know, at the individual level, think of all the little things that you just let go because you know, you just don't have the time or the bandwidth or the mental space to deal with them.
Bridget McCormack: I have an insurance issue. I just keep moving it to the bottom of my pile because I don't feel like having to fight and I don't even want to figure out where to log in and what to write, and then what do I do from there and I'm a sophisticated user of insurance service products and dispute resolution services, and yet I don't resolve disputes. And so I do think that because we're building services that really meet that enormous need that right now is kind of unmet. I think that's captured the attention of many. It's also the case that for some reason I think the public is really interested in what does it mean if judges use AI and does that, you know, somehow make judges less judgey. That's probably not how they're putting it, but you know what I mean. There are folks who think if a judge uses AI somehow, it's no longer, we've kind of given up on the sanctity of the court system.
Bridget McCormack: I happen to think it might be a way for the court system to grow trust in what it's doing because with dispute resolution outside of courts. There are all kinds of ways they can scale solutions for the people who need them that will make a big difference. But that's not what you hear about right? You hear about those two federal judges who released opinions with hallucinations in them.You know, how terrible is that? Meanwhile, the rate of errors that we see in any normal judicial system is so high and we're apparently completely comfortable with that. Just look at the number of cases that get reversed every year by Intermediate Appellate Courts or State Supreme Courts or the US Supreme Court, and you'll see that apparently we're quite comfortable with human judges making all kinds of errors that we're fine with but the fact that two federal judges released opinions, the more
Zach Abramowitz: human process.
Bridget McCormack: Yes, exactly.
Zach Abramowitz: I think there's so much to unpack here, but I think that there's different processes that we really require a human for either all or part of it. Only because that's the way it has been up until this point. But that's also because we operated in a world where intelligence and reasoning was not available on demand.
Bridget McCormack: Yes
Zach Abramowitz: If you read about the early days of the American Arbitration Association, there were human arbitrators that took on these cases because they needed an unbiased, neutral party. As far as I've read a lot of the arbitrators, if not all of them, were taking these cases pro bono.
Bridget McCormack: Yes
Zach Abramowitz: It was not seen as something that you were necessarily going to pay for and you used a human because that was the available source of intelligence. What else would you have used? We're going to have to go through and do some serious societal unbundling and ask ourselves what things do we really want a human doing? What things have we used a human up until now because that was the only available source of reasoning and intelligence?
If you think about dispute resolution, and you made this point in the article in the Verge. In a lot of ways, an AI is not just able to replicate the human arbitrator in many ways it is potentially a more powerful dispute resolution force than is a human. Especially when the alternative in many of the cases you were highlighting before might be no dispute resolution at all. So it's for the cases that we're not ready to move off of humans for, that's where society is. But for every case that didn't have an appropriately priced intelligence, now we have that. I think that's a really important point. I think it's going to apply within the courts and within dispute resolution and probably in every other vertical where people have to unpack and say what do you really want a human doing? Maybe the difference between a nurse, where you probably want an active human versus, in a certain world you might imagine reading radiology scans might be able to be better by an AI.
Bridget McCormack: Yes. A hundred years ago arbitration was an innovation, there was one way everybody resolved disputes. It was going to court. You went to court with a lawyer. There was a set of rules and processes that had been built 150 years before that. And that was working okay. Then it was working less well, and it was working less well with larger disputes like World Wars. The AAA, very audacious, founder Frances Kellor, who decided we needed to be able to figure out new ways of resolving disputes and the most important element of that would be giving people at the individual level control over those disputes, agency over those disputes. I see what AI is offering now as just like supersizing that opportunity. Because if you, as an individual, might think that for certain disputes in your life you still want a human, you want it to be in a public courtroom. You know that that's right for you, depending on the dispute, but there might be others that if you could get them resolved on Saturday morning on your smartphone and move on that you'd be thrilled to be able to do that. And so it really is just giving people, businesses, governments, communities, more options for moving through disputes. Disputes are unpleasant and they can be devastating, but they don't have to be, if you can figure out constructive ways of moving parties through them.
Bridget McCormack: that's all of a sudden possible at scale if you can build, like we have, although, for one small use case so far, but a governed, transparent, trusted new option for parties when that new option makes sense. We're just expanding the toolkit, and it's about time.
Zach Abramowitz: Yes. You mentioned this in the article, and I don't remember if it was in the Verge or the Wall Street Journal, but, you're not advocating that people go and resolve binding debate, binding disputes, the type that you might theoretically bring to a court. You're not telling them to do that in ChatGPT, right? That's why you've invested as much as you have in the AI Arbitrator.It needs to have the force and stamp of the AAA behind it. No one would advocate for just using ChatGPT for some of those disputes. On the one hand, the model on the other hand, we have to treat disputes with the respect they deserve as well. And that means a reasoned AI Arbitrator, but one that's based on the know-how and intelligence of the AAA.
Bridget McCormack: Yes, I think you could throw a dispute into ChatGPT and would probably do a decent job in most cases right now. That's just where the models are, but you wouldn't have any way of knowing how it did, what it did, why you should trust it. What it was trained on. So I think that's the value add of the AAA building and agentic system. Each agent is trained on our expertise and our previous decisions and handbooks that we've built from those is the AI dispute resolution we're building will show its work and grow trust. People will see how it works and they'll be able to decide that that's a good option for some disputes. I think there are some disputes that are probably never going to make sense to put into AI dispute resolution, CRI cases the government brings against its citizens and cases its citizens bring against it. I think those cases should happen in public courtrooms, with reported decisions but private disputes between parties who want to move forward with their lives and want to do that effectively and efficiently. They should have lots more options and now all of a sudden they can. I think that's what's exciting. I think everybody feels that, and that might be why there's been interest from the media, beyond legal media, in what we're up to.
Zach Abramowitz: I wanted to also emphasize another point I use the American Arbitration Association, at times as a kind of shining example of what to do. I think there's several areas where the American Arbitration Association is a true outlier. And the one I've mentioned to you before is that in most companies, AI is employee led, not employer led, meaning the employees of a company are using AI long before that company actually purchases AI tools. They're just sneaking their AI into workWe've talked about the surveys that seem to show that pattern over and over again, but the AAA is not that way, At The American Arbitration Association. The AI was very much led by the company, making it available to their employees at the earliest phases.
So I think there are a lot of ways in which you've been outliers, but I think, one of the ones that I like to highlight and I've been highlighting recently, is the ROI of the brand is that it seems right now, and I would say think the coverage is a kind of symbol of this or, or at least a symptom.Is that when you invest seriously in AI in a way that is credible to the outside world, not just as innovation theater, but as we're all in and we mean this, that shines through and people were able to pick up on it, whether it's your users or the ecosystem at large, you know people will come over to me at conferences and say, is The American Arbitration Association hiring? I'm dying to work there. This is a new experience. Something that I didn't experience pre 2023. I didn't get people coming over and angling for a role at the American Arbitration Association because it was a top tech company. I think there's been a lot of real street cred developed. Can you talk about the ROI, for the brand and how it sends that signal to your users?
Bridget McCormack: Yes, and I think you see that in the larger marketplace, right? You probably saw Walmart's earnings last week. They've increased their market cap so significantly. And they don't own the AI stack. They're not building a large language model. They're figuring out how to use it intelligently to benefit their users. So there's natural language Q&A that turns out to be pretty effective and users like it. They're using it in contracting, they're doing agentic B2B contracts and saving money there. I think when the world sees that you're trying to figure out how to make this technology work for the users not to make their experience worse, but to make their experience better, which the technology absolutely can do if done right.
Bridget McCormack: I think you get this brand boost. That's kind of independent of the more technical ROI. I think everybody's trying to figure out how to use it. The businesses and the organizations, legal and not legal that are. Showing their users that they're really serious about building new services or building new functionality to their current services that make their users experience better are the ones that are getting that boost. I count us among those. because our view is that this is simply a way to scale what we've been doing for a hundred years. It's an indirect service of our mission that we can find new ways to give people options for resolving disputes.
Zach Abramowitz: So tying AI into mission and DNA, this is what we're supposed to do as an organization and here's why that means AI, that's a critical component.
Bridget McCormack: Yes exactly
Zach Abramowitz: Bridget, we're looking forward to recording more podcasts with you during legal week, and I'm looking forward to the panel on Tuesday. So if you're there, make sure to stop and say hello to Bridget, but I'll also be in the audience, so make sure to stop and say hello to me as well, and we look forward to seeing you at the next episode of The AAAi podcast.