Bridget McCormack on Court Modernization, Access to Justice and AI in Dispute Resolution

I'm excited to be here, excited for the conversation. Thanks for having me.

Yeah, I don't think it's I don't think there was a time when my thinking changed about the role. That's that I think it's it's something about the role that very few people understand. State Supreme Courts are charged. In most state constitutions with administrative oversight of the courts of their state. And so state Supreme Courts approach that role differently. You know, some delegate a lot of it to a judicial counsel or a state court administrator and his or her team. Others are rolling up their sleeves and really trying to figure out how to make the. The state court system work better for the people who need it. And the Michigan Supreme Court is in that latter category. It was one of the reasons I was attracted to the job in the first place.

I think the decisional part of the job is obviously critical and pretty important to all the rest of it. But the opportunity and the responsibility to figure out how a justice system can serve the people of a state better. Is a much harder problem and a really, really important problem. So that part of the job actually attracted me to it in the first place.

I have zero things on that list, Brendan. Sorry. Sorry to disappoint. I think that if you approach an institution as important as a branch of government. That is delivering justice to people who need it. And also extracting it from people who need to have it extracted from. And you think that there are things that can never change, you're going to have a hard time figuring out how to serve the people you're serving. So there may be things that take a long time to change. There may be things that. You know, after due consideration, people like the way we've been doing it for 250 years. But I think if you have anything on a list where you say that can never change, it's going to be you're going to have a hard time, you know, negotiating a complicated, fast changing future.

I don't think that Michigan is different from any other state, but I do think that there are state level practitioner concerns that are. Different than most of the litigation in federal courts. And I think it all comes down to volume. I mean, the state courts are handling just significantly more cases. And whether the particular lawyers are or not, whether a particular law firm or a particular lawyer is handling more cases than her counterpart in a, you know, in federal court isn't really the point. But the courts themselves are just dealing with such high volume that it's harder for a court to focus on the specific hard questions. And that's not only in discovery and eDiscovery, it's across the board. But given how. How rapidly discovery has been changing.

And I mean by that, like how much more, you know, information there is that has to be managed in the course of a of a of litigation that it just it just presents unique challenges to state court judges that I think are true. Not only that, that's not truly for Michigan, but most state courts.

Yeah, I mean, we kind of went about the process. And with, again, with an anything is on the table mindset, you know, so we engaged practitioners who, you know, who spend most of their time on one side of the V or the other. We engaged professionals in the eDiscovery space and we engaged trial judges who were. Really struggling with managing the issues before them to learn from them, like what they thought would work well in Michigan, in our state courts, and whether that was what the federal courts were moving to or something different. We were completely open to it. But I would say the court in particular had no preconceived notions about what the right answer was. It was really driven by. The court users, you know, eDiscovery was one place where it's it was hard to hear from the lawyers clients.

You know what I mean? I actually believe that the people that the court serves are the parties more so than the lawyers and certainly more so than the judges. And so in lots of court reform issues, we tried to figure out a way to hear from the actual people who are impacted by. A process change or a process update. That was harder in a discovery. And so we really had to rely on lawyers' representations about what their clients wanted, what their clients cared about, and what was going to be most beneficial to their clients. And I think it works in that context. If you guys have different ideas about that, I'm interested, though. I always want to hear it.

Yeah.

Yeah, maybe more so than attorneys, because as I'm sure you're aware in state court cases, I think the National Center's most recent statistics have it at 75 percent or higher. At least one party is not represented by a lawyer. So most people in state court systems are navigating their civil justice problems, not criminal, but civil justice problems without lawyers. So you have to figure out what works for them because we're never going to lawyer our way out of that problem. So figuring out what works for people who are trying to figure out how to get a little bit of justice is pretty important. We did a lot of. Town halls before the live town halls, before that, before the pandemic shut us down, we had to move to the move to this platform, to the Zoom platform when we were launching a justice for all commission, because we really needed to understand from the people in there, in the different communities across the state, what it was that they wanted and needed from their local justice.

System, their local court system. Similar process with the Jail and Pretrial Task Force, which was actually a multi-branch, multi-stakeholder task force, but I co-chaired it with the lieutenant governor, and we did town halls around the state where we literally heard from people who had been impacted by the jail and pretrial policies. I don't know any other way to do justice system reform than to hear directly from the people who are impacted.

Yeah, in some ways, Pandemic gave us all this data that would have taken us forever to otherwise get in courts about remote proceedings and who they work for and why. Michigan was lucky. In December of 2019. We purchased Zoom licenses for every judge in the state, and I wish I could say that's because we were prescient. That's not true. Michigan's just a huge state, and it made sense to us for a lawyer to be able to appear in Traverse City in the morning and Detroit in the afternoon. So it seemed like a good idea. That didn't mean that anybody was using them by March of 2020, but we had the functionality. So we were able to very quickly stand up remote courtrooms around the state and spend our sort of innovation hours and budget on other things.

Like we built live streaming for the public to be able to watch those court proceedings. And we threw up a website where the public could literally click on a county and then on the right courtroom and just watch what was happening. And then a million other things from there. But back to the really important question you asked, and I appreciate it. We ended up collecting a whole bunch of data. Lots of it was just quantitative. Like we knew after some number of months and then some number of years that the default rates, especially among self-represented parties when they had a remote option and they were they were significantly lower. In retrospect, I like to say, duh, because, you know, there's a lot of barriers to getting to a physical courthouse like.

Transportation. In Michigan, we don't have public, we don't really do public transportation so much. We believe in cars. And cars are expensive, a lot more expensive than a phone. And sick kids, sick parents, disability, there are lots of reasons why the barriers to getting to a physical courthouse are significant. A job, a new job. I mean, I watched a lot of online proceedings during COVID. And I can't tell you how many times I saw a parent or a criminal defendant who had gotten a new job, but was able to make his court appearance in his break room and therefore not have a warrant out for his arrest. There was no way he could have gotten a day off from work, but he really wanted to also make his court appearance.

And he was able to do that. And, you know, again, retrospect, of course, that all seems obvious. But we all of a sudden had data on all of that. In addition to that, we had some qualitative data that is, I think, as valuable. We heard repeatedly from judges who oversaw cases with children as either parties or impacted by their parents' decisions. Repeatedly, judges said, I'm getting just higher quality information from kids when I can talk to them in a video context. You know, one judge told me, you know, this one girl would never, she had a very hard time talking in a courtroom, which I understand it's, the courtrooms are scary even for lawyers, right? They're just like scary places. Imagine putting a kid in there and like trying to like have them like, you know.

Tell you something that's hard to talk about. But in her in her room at home, she wanted to show the judge, you know, her artwork and show her homework. And so the judges were getting like better quality information out of certain kinds of people. Legal services lawyers who, you know, are limited by, you know, being human beings and having to get from, you know, courtroom courthouse to courthouse in, say, an eviction like docket. We're able to do I think they the Southeast Michigan Legal Services folks said about seven times more cases when they had the opportunity to appear remotely. So it's like across the board, win to have a remote option for, you know, most proceedings, obviously for people who don't want to appear remotely would prefer to be in court.

You have to have that option, too. But the idea that knowing what we know now that you wouldn't offer. Remote options in almost all cases, it feels pretty user hostile, if you ask me. And I know that many states have gone back to requiring in-person attendance for everything, and Michigan's not one of them. And I think that's a mistake. You know, the courts are for the people. They're not for the judges. For the most part, the people don't even really want the judges. Frankly, they don't want the lawyers either. They want to solve their problems. We are, you know, as Richard Susskind says about the drills, you know, they want holes in the wall. They don't want to buy drills. They have to buy a drill to get holes in the wall.

Well, they have to put up with lawyers and judges to solve their problems. They're not really in the market for us.

So I think the data point you're looking for is a recent paper by an MIT scholar and a USC scholar. I'm not going to pull up right now. They looked at court data, but only federal court data from the change from 2024 to 2025. And sort of the 20 years before that, the. The filings were basically steady. You know, they really didn't they really didn't change. The self-represented parties filed the same amount, the same number of kind of cases year to year. And in 2025, the increase was maybe 14 percent of significant. Don't quote me on that. I can look it up. But, you know, it was a significant jump. It's also the case, by the way, that same paper that the number of pleadings filed by lawyers has significantly increased.

Lawyers also are able to use AI to file a whole bunch more motions. So courts are seeing increased filings not only from people who don't have lawyers, but also from lawyers. The most interesting finding in that paper is that the number of cases that were dismissed for want, you know, for not being able to state a claim or, you know, were improperly like not shouldn't have been in court at all remained the same as in those prior years. So even with increased filings, we were seeing that, you know, what we know is a lot of those needed to be filed. Right. These were like valid claims. My guess is that the state court data is going to look significantly. More more robust. We're going to see even more increased filings in state courts because state courts is basically where most people go to to get justice.

You know, 98 percent of civil and criminal cases are adjudicated in state courts, not federal courts. And I have anecdotally talked to friends who sit on state Supreme Courts and my friend Beth Clement, who's the CEO at the National Center for State Courts. And they say they are that the filings are just off the charts. But if they are cases that need to be there, I think that we should all view this as like a really good news, right? Like a very good news story about AI. Now, does that mean that we need to figure out how to manage all of those disputes? Because it's not like we were managing great before. You know, the courts have been burdened for a few decades now. You know, the courts kind of outgrew the...

They were designed 250 years ago for a system where every single person had a lawyer and every matter had plenty of time for a judge to consider it. That hasn't been true for half a century. And we haven't updated the operating system to respond to it. So, yes, it probably means we need to figure out how to update the operating system. But I think it's. Clearly good news.

Yeah, I mean, AI offers incredible opportunities for more choices for parties that have disputes that need resolving. Of course, those solutions have to be built with transparency, with the ability to be audited for bias and accuracy. And they can be systems that are both assistive to courts. And also in cases where the parties want it, they can supplement, you know, dispute resolution options that that courts already provide. And you can imagine lots of areas where parties might prefer that. I mean, we see it in arbitration already. And so I think it's you know, it's been a long time since we've had a new kind of set of options for resolving disputes. We now have a technology that can do that. Of course, like the things that people worry about are problems that need to be solved.

But we've been pretty patient with those problems in an analog system for a long time. Like if one more person talks to me about hallucinations and I have to remind them about human hallucinations. Including human judicial hallucinations, I'm going to lose my mind. I mean, you know, we have an entire appellate system built to respond to some of the hallucinations that judges make. As you all know, very few of those hallucinations qualify for appellate review, right? We have all these rules about which ones we'll even look at. But we've tolerated human hallucinations forever. We now have a technology that we can require to show its work along the way. If that's what you want, you know, we can you can require that. You can build in the kinds of checking that you might want.

You can build in the responsibility for the technology to tell the parties how it understands. What they think their complaints are. There is so much upside potential to building new solutions and solutions that some parties will want in some cases. I think there will be lots of other cases where parties won't want that. They should be able to have whatever they want.

Yeah, I mean, I definitely wasn't I didn't think I could show up and know what needed changing, if anything. You know, I always show up at a new at a new place and learn a lot before I before I form thoughts about how I can contribute to to where it might go. The AAA was exciting because, you know, it's a 100 year old institution that administers over half a million cases every year around the world. So it's really. The next largest dispute resolution system, once you leave, if you leave a state Supreme Court, you know, in Michigan, we resolved, I think, between three and four million cases a year. The AAA is close, right? It's it's close. It's more than probably some smaller state court systems.

And we have an awful lot of cases where self-represented parties, where parties represent themselves often on both sides of the V. We have a large self-represented party commercial caseload. And construction caseload. And so it really is an opportunity to think about dispute resolution and also to be able to innovate and take advantage of technology in a way that's harder in a public justice system. I mean, the courts have to rely on another branch of government for their funding. You have to remember that all of the judges across the state are separately elected. And so you might have a new court rule on remote proceedings and some of them might not think they have to follow it. So, you know, it's a it's a it's a much harder change management environment for dispute resolution than.

A 100-year-old nonprofit that gets to figure out what works and then do more of that, right? If we're doing well, we can continue to invest in R&D and build new processes and new services that work. So it's kind of a wonderful opportunity to take lots of ideas about street resolution. And and and build and start building a future at a time when, you know, the world was changing. Right. You know, I took the job. I started the job in February of 2023. And, you know, Sam Altman changed our lives like two months before that. So the timing was exciting.

Yeah, it's a great question. You know, we've we've been basically building point solutions and then enterprise solutions and then brand new products and services with AI for three years now. We armed everybody across the AAA with enterprise licenses to chat GPT. We took volunteers from our panel of neutrals to arm them with a lot of the legal tools that were on the market. We got them subscriptions to those. So we could just experiment and figure out what we could build that would either improve our operations, like something that would make what we do internally better, or that would improve the services we provide to the users. Self-represented parties are a big target there because we have a lot of information that is, we think, helpful to them.

But, you know, when it's like scattered across a website and you are not, you know, you're reading these rules and you don't even really understand when they kick in and you have nobody to talk to, that's complicated. And obviously, AI gives you an opportunity to build natural language interfaces across all that information. That's really tremendous. I don't know if that explains why we have significantly more self-represented parties filing with us this year than we have ever before. But but we do. Another, like, so we basically, we took the approach that, like, let's take all ideas. We have an innovation platform that connects everybody across the AAA. We're scattered all over the country, and we have a small team in Singapore. We upload to a shared platform called Bright Idea.

Then when a new idea is uploaded, people workshop it. It goes from there into an innovation pipeline, and then we have to do a little bit of... Thinking about whether we can afford to build it, whether we have the capacity to build it. And then we built a lot of stuff. Some of those things have been fantastic for users or just internally. Some not. And those we learned just as much from. When something didn't work, we were like, oh, okay. So that's not useful for this reason or that reason. That helped us. That helped us figure out how to place some bigger bets. And the bigger bets are now either... In the wild or, you know, about to be in the wild, we released the world's first arbitrator in November, which is really a series of agents that operate across the arbitration process.

There's, you know, at the front end, a series of agents that take the information from the parties, self-represented or lawyers, take like their complaints, their evidence, their any legal support they think that supports their position. And then those agents take everything they have understood and organize it and tell it back to the parties. They come back to the parties and they say, this is my understanding of what your complaints are and what evidence you think supports each element of each claim and what legal support you believe follows your argument. Did I get it right? And the parties then get to say yes or almost. You know, there's this one thing I want you to add to that. You know, this evidence that you said supports that claim.

It also supports this claim. And then the agents like incorporate it again and go back to them again and say, OK, here's here's what I think you're telling me. Do I have it right? Imagine it's like procedural justice, procedural fairness, like. On display. And I have to tell you, I did not appreciate that at the front end of the bill. I knew that we should build it because I knew that bringing down the time and the cost and bringing up the reliability of an arbitration process for certain parties was important, was going to matter, was going to make a difference for people to have that option. Again, they should choose. They should have whatever they want. They should pay a human and take a really long time if they want, or they should have that process.

I did not appreciate that. The unbelievable bonus to procedural fairness, which, as you know, there's a robust academic literature about, it's really important in courts. We spend a lot of time, you know, in judicial training because on how important it is to make sure that parties feel heard and understood. There's, you know, the data shows that they are. Far more likely to accept an adverse outcome to their position. They will take bad news. They will take a loss and still grow trust in the process if they feel like they were heard and understood. That's really important in courts because today's parties are tomorrow's jurors and tomorrow's witnesses, and we need the public's participation in a public justice system for it to function. And if people lose confidence because their experience is one where they don't believe they were heard or understood, that causes more undermining to the rule of law than anything else I can think of.

Well, now we have a magic language calculator or a series of them that can literally go back and forth as many times as the parties need to make sure. They feel heard and understood before those agents then pass it on to the next set of agents that reason and draft and then pass it on to a final group where there's a human arbitrator in the mix. We have a human arbitrator in the mix. So those are like bigger bets, and we're now training that for other use types. We started with documents-only construction cases, and we're building out other use cases and case types on that platform. Those only came because we spent a few years trying to see what we could do, what we were capable of doing ourselves, what we needed to do with a partner, what people wanted, what was going to be additive to the services we already offered.

But I will tell you, this technology is moving so fast that if we talk again in six weeks, I promise you I'll have more ideas. We'll have new things that we know we have to put in the... in the pipeline.

Yeah. I'm a little bit of a broken record. I'm going to tell you what I think personally, but I don't think my view should count for a whole lot.

No, I think parties who need disputes resolved should be able to choose. But let me answer your question, and then I'll give you all my caveats. Cases against the government, so criminal cases where the government is, you know, is trying to take away someone's liberty or constitutional civil claims against the government or other civil claims against the government. Have to have a person who represents the government. Making the final decision and looking somebody in the white of their eyes and telling them what that final decision is that they passed on it and they believe in it. Civil disputes, especially disputes where the parties have drafted a contract about how they resolve their disputes. I'm not sure why it's my business. I think they should have lots more choices than they have today.

Better choices than they have today. And especially now that we can build better choices with this technology where you could have. A little bit of an assist from it. I can't imagine in some number of years that anybody would pay human arbitrators to do some of the building the timeline or some of the work that the technology can already do better than humans. And we know that today. And why would you pay humans when the technology could do that? And they might still want to be able to have an argument. In front of a human being and have a human being pass on the final decision. And there might be some parties who say, for all of our supplier disputes, we want those under some number of dollars in dispute.

We think those can go completely to an automated process. I think there'll be other disputes forever. Where parties will want a human from the beginning to the end. And those won't necessarily be just like that, the company civil cases. There'll also be cases where parties need to have a continuing relationship. So family court, you know, family law cases is like an obvious one. When the where there are kids involved, you know, where the parties have to have a continuing relationship. And it really matters that relationship comes through conflict well. So my guess is a human being will continue to be an important part of helping those parties through that conflict. I could be wrong, though. Like, you know, again, talk to me in six weeks.

We'll see what the technology is doing next. Maybe, you know, maybe it's better than humans at that, but I doubt it. And there are lots of employment relationships like that as well. You know, there are lots of conflicts in. In business where the parties have to continue to have a relationship. And a human navigating that with them will probably continue to be important. Having said all of that, I really think it's very main character of me or any judge or arbitrator to say, these are the cases that... humans have to, you know, be involved in? Like, why? I mean, unless I'm a party, then I'll tell you if I want a human involved in it. Otherwise, like, why? If we have ways to give people lots more options to resolve their disputes, and they choose the one that works for them, and they get through the dispute and get to the other side in a better way than any other option, why isn't that the right answer?

What do the parties want?

Yeah, yeah, yeah. I mean, that's actually one of the upsides of being able to see the work that the agents produce and then see any changes that a human arbitrator makes. And we did that throughout the design sprints to build the product. We had human arbitrators from our panel in every design sprint giving feedback on the agent's work. So, yeah, that's that's that's pretty critical. I always say, like, imagine if a human judge had to show her work, you know, had to literally show exactly how she got from this to that to that. You know, I mean, judges don't have time. I don't mean to this is not a critique of, you know, of judges. But but imagine if they had to go back to the parties first and say, here's my understanding of your case.

I believe that these are your claims and this is the evidence and this is the way you think the law fits your claims. And, you know, you party B, this is what I understand to be your claims. I mean, even imagine if judges had to do that, it would be I mean, again, there's no way they have time for it. But it might. But it might be interesting. I mean, I would love to see that, like, if you could give me, like, a printout of, like, some judge's brain in a decision that I don't understand, I would love to see that printout. I want to understand it.

Yeah, it's such a good question. So, yeah, six weeks is about the timeline. I do think and I think it's actually I'm going to do my best to answer your question, but I think our human brains can't do it. We can we are really good at what really good. We're somewhat good at projecting in a linear way. We. Are not capable of projecting in an exponential way, and we are living in an exponential timeline. And the impacts on the legal profession and the justice system's little corner of that legal profession are coming from all sides that aren't in our control at all. So the number of filings that we talked about is off the charts. That also means that the access to information is just growing, which is tremendous, frankly.

The discovery we're going to see across science and medicine and engineering is so exciting, but it's going to cause all kinds of really important legal questions over the next weeks, months, and years. We are starting to see agents forming contracts. You know, agents are forming contracts with other agents and not just like B2C, B2B. And we have no legal framework for what what law governs. And I've given it a lot of thought and I don't think that the e-commerce law that's grown up for the last 20 years fits. So we have entirely new areas of law that are just going to have to be developed. I think all of that can be overwhelming. But we're moving from a world where legal information and access to it was scarce and therefore guarded by the guilt to a world of abundance where legal information is going to be abundant.

That's good. You know, I mean, the idea that we have this public good, the rule of law that has, you know. That is responsible for the greatest economy in the world, right? The, you know, and lots of other, lots of other good things, but is guarded and not accessed by most people has held us all back. It's not, it hasn't been, it hasn't been good for communities. It hasn't been good for the economy. And so moving to a world of abundance is in my view far and away. So there's a lot more good than bad, but it's going to be a little messy along the way to figure out how it all comes together. But I am, at the end of the day, very optimistic.

Yes. Thanks for having me.