How AI Is Quietly Transforming the Courts

Episode Notes

Transcript

Matt Davis [00:00:00 – 00:00:13]

So the other interesting use case is the Arizona Supreme Court. They’re using AI avatars to explain the opinions from the high court in plain language.

Phillip Vaden [00:00:14 – 00:00:15]

Really? That is interesting.

Matt Davis [00:00:17 – 00:00:38]

They want, they want the public to understand these rulings. And you know as well as I that, that, that they’re very legal. They can be very hard to read for the general public. And I think that’s a really neat use case and one that is really beneficial and right in the heart of what AI is very good at.

Phillip Vaden [00:00:39 – 00:01:38]

When it comes to the intersection of AI and the law, most conversations center around legal professionals. But what about the courts? That might be where some of the most interesting challenges and opportunities do and will exist. Coming up, we’re talking with Matt Davis, Vice President of Product Development at Tyler Technologies. Matt has over 20 years of experience in public sector software with particular expertise in courts and justice technology. Matt’s lending us that expertise to talk about where courts stand with AI today and where they’re headed next. Given the wide universe of possibilities, it’s a good thing we’ve got some Star Trek and Battlestar Galactica references thrown in along the way. I’m Phillip Vaden, partner manager at InfoTrack, And this is the AI Sidebar. Matt, thank you so much for joining us today on the AI Sidebar.

Phillip Vaden [00:01:38 – 00:02:27]

I appreciate you being here and taking the time. You, uh, you have all of the knowledge that we want about what’s going on in the court space. And, um, you know, we’re going to kind of look behind the curtain here. We, we typically interview people on the, uh, the, the lawyer side, the paralegal side of things. And so we’re really excited to get a peek behind the curtain of what’s going on in the court space, specifically as it relates to artificial intelligence and what all they’ve got going on. So thank you for letting us pick your brain about this topic. So let’s start with the court space and dive in there. And in what direction do you see the courts moving in and what kinds of technology are they looking for that Tyler is looking to address and solve for them?

Matt Davis [00:02:28 – 00:03:09]

Well, sure. So one, I’m really happy to be here, Phillip. Thank you for having me. What we’re seeing in the courts right now is really just what I would say is the tip of the iceberg when it comes to AI and what you can do. If you look at what the courts are doing right now, we see 3 major use cases. The first one that we see them using it is really for chatbots. And this can be a variety of things, right? It can be external chatbots for the public. It can be internal ones for their users.

Matt Davis [00:03:11 – 00:04:04]

So if you think about Internal ones. The examples that we see there are really about their internal processes and their staff and their training. You know, these government agencies, they don’t always have the most robust of budgets, and they’ve got to do the best that they can with what they’ve got. And that often means that they have high turnover. And so one of the best ways that they’re using AI right now is to ensure that their, their new staff is, is well trained, up to speed, and can operate in the way that they expect, right? That’s not a direct impact on the public, but it helps to, to ensure the ship is running. And is effective, right?

Phillip Vaden [00:04:05 – 00:04:05]

Yeah.

Matt Davis [00:04:05 – 00:04:18]

Yeah. We see that in locations like Orange County, California, and their Superior Court. They have an internal chatbot that they’ve built that they’ve been using for a while now.

Phillip Vaden [00:04:19 – 00:04:20]

Very cool.

Matt Davis [00:04:20 – 00:05:03]

On the external end of the house, we see like the, like the Alaska court system, they built one for For, or I’m helping SRLs with estate and with probate so they can help navigate the legal system there. That one’s a little more interesting, obviously, because they worry about things like UPL and hallucination, right? So they’ve got to be very careful that those systems are well trained on only Internal documentation.

Phillip Vaden [00:05:04 – 00:05:04]

Okay.

Matt Davis [00:05:04 – 00:05:16]

So that’s the biggest thing that we’re seeing them doing right now. There’s a couple of other areas where they’re using AI, but any thoughts or questions on the chatbot area?

Phillip Vaden [00:05:17 – 00:05:44]

Yeah, on the, on the chatbot. So, so this is, um, uh, in, in, for this setup for the, the courts, do they have, I guess they, they’ve got manuals that they, they load in and, and it’s basically like, hey, we’re gonna expedite getting people up to speed, but then they’ve, hey, if you have a question and everybody’s busy, you can hit this chatbot and get your answer really, really fast. So is that, that’s the general setup there for that internal chatbot? That’s great.

Matt Davis [00:05:44 – 00:06:23]

Yeah, that’s exactly right. I think a lot of us right now have realized that maybe AI, these, you know, large language models that They have knowledge about almost anything in the world, right? And you can open up OpenAI or Anthropic and ask it anything about any subject and get an answer. And the availability of just broad knowledge is almost a commodity now, right? And so this is the same thing, but they’re loading it up with their internal knowledge, their specific knowledge. about the business of the courts.

Phillip Vaden [00:06:24 – 00:07:04]

I’m going to ask a further peek behind the curtain on this, because one of the things we’re seeing in sort of more broadly in the AI space is people have, we’ve seen some instances of token overrun where all of a sudden people have accidentally spent hundreds of millions of dollars on tokens. So in these scenarios, what does the setup look like? You don’t have to get into like the nitty-gritty details, but what does a setup look like for reducing the number of tokens? Is it just a cloud-hosted environment that’s, hey, we’ve got this model that’s doing this thing. We don’t need a frontier model to do this kind of work. Is, I guess, what does that look like to prevent the token overruns?

Matt Davis [00:07:05 – 00:07:41]

Well, I think that sounds right for this use case. I don’t know the specifics of the way each of them have implemented this, but broadly speaking, summarization of information or explaining of knowledge wouldn’t require the latest frontier models and could be used with a lower-end model. I would, I don’t know for sure, but I would wager that they’re doing their own hosting of model or using a lower-cost model like a Haiku out of Anthropic as an example.

Phillip Vaden [00:07:42 – 00:09:30]

Yeah. And so that gets into another area of trust. And this has seemed to be a broader problem both on the private side with law firms, And then obviously with the public sector side as well, where there’s a lot of people out there that are looking at Anthropic and OpenAI, amazing things that they’re doing, both companies, but the level of trust is not very high in like, what are you doing with the data? We don’t really know exactly what’s going out there in the world. And so Obviously, on the private sector side, it’s attorneys and they, hey, am I getting rid of my attorney-client privilege by entering this specific prompt? Am I injecting private customer or client data into the system with this prompt? And then on the government side, it’s the same thing, because eventually a lot of that information is going to go into the court systems. And so if you have, for example, like a divorce with children case, You don’t want the kids’ names getting out there. And so on that front, do you feel like if there is a setup that is built where people are able to have more contained AI, basically their own internal walled-off artificial intelligence sandbox, basically where it’s like, hey, only this court, only the state of Texas or Dallas County or whatever, all of their information is in this contained environment running on X model, Haiku, Sonnet, whatever, that that will alleviate a lot of the concern that the court has about more fully adopting artificial intelligence?

Matt Davis [00:09:32 – 00:10:39]

That’s a good question, Phillip. You know, I want to begin with, you know, saying that you’re exactly right. And what you describe is the very foundation of the way we we look at AI. The foundation of the features that we’re building are really based on it being in a sandbox environment where there’s the full governance, there’s the trust, there’s the bias control. We don’t use any of the customer’s data for any reason outside of their purposes. We don’t, to training of models. We don’t do any of that, right? And the LLMs that we use are ones that are guaranteed to be, to be a closed-off sandbox where, where that information is not able to be used for anything else. And while I think that will help.

Matt Davis [00:10:39 – 00:11:45]

I do think there’s a reality that the courts are wary of their information. You know, they’re the custodians of the data. That is their responsibility, is being the custodians of the data. There are, you know, some courts out there that are even hesitant to put that data online on a public website, right? And so their level of comfort with AI will be even lower. So there will be, you know, some ports out there that will be more rapid adopters, more comfortable with AI, understanding what they’ve got in their contract with their with their vendor and the limitations and the controls around that data. But there will be other courts that will take longer because they just want to be very careful with their responsibility as the, as the custodians of that data.

Phillip Vaden [00:11:45 – 00:12:28]

Yeah, absolutely. So one, one thing that’s, that’s come up again on the private sector side that I’m— and some of this, I mean, I think the chatbots are a great example of of augmenting the clerks, the reviewing clerks at the courthouse. So are there any other things that courts are asking for or adopting already that augment their workers and allow them to, you know, hey, I can do more with less because it’s really tough to get people in. It’s tough to keep people here because obviously, you know, the pay in the government sector isn’t always the the best, but they do as much as they can with what they’ve got budget-wise. So I’m curious about that.

Matt Davis [00:12:29 – 00:14:22]

Yeah, you’re right on. That’s the next category of where AI is being used is around automation, right? And we see this a lot with our own automation solutions that we call document automation. But the idea is that you, you use AI to automate low-level tasks, things that are repetitive, things that are redundant, but have historically required a human, like e-file review is a really excellent example of this, right? The public and the, the attorneys really want rapid review of their, of their filings when they, when they submit them. But historically, there’s been a dependency on the staff to perform that task. Using AI, you can automate the rules, the compliance rules, the review of the data, understanding that everything’s there as it should be, and maybe accept that filing automatically if there’s a level of confidence That meets a certain threshold. Broadly speaking, automation is a realm that the courts are more comfortable with, right? Where they shy away, and rightly so, is in the realm of decision-making or impacting outcomes, right? They don’t want AI anywhere near that. And that’s— I think that makes a lot of sense, right? But automation in their day-to-day work, especially for the low-level repetitive tasks, is a major use case, I would say.

Phillip Vaden [00:14:23 – 00:14:46]

Yeah. So a little peek behind there. So Matt and I, we worked together for 10 years. Matt’s been at Tyler for over 20. 10 of that was with me. I’m so sorry for that decade. No, but the, But yeah, so for the courts, it is, you have, there’s so much work to do. There’s so much work to, for the clerk’s offices to take on.

Phillip Vaden [00:14:46 – 00:15:53]

And a lot of it is things, you know, the shift from paper, which in some places, I mean, you go to the East Coast and there’s courts, I know that visiting with the clerk in Chatham County several years ago, and they have court records going back to the 1700s. And so managing that and everything is a big, big undertaking and there’s so much work that can be done by the staff if you can take some of these highly repetitive tasks off their plate. And it could be things like going in and digitizing all of these older records so that they don’t get lost to floods and fires and hurricanes and everything else. Because if a bunch of paper sits in a warehouse, it just feels like over a 100-year time period, there, there just might be a fire. Um, so I think that that’s, that’s great to, to free up the, their time for, for higher leverage tasks that, uh, that do require, uh, uh, humans absolutely all the time. So that’s awesome. And then decision-making you mentioned. So let’s talk about the, the, the judges for just a minute.

Phillip Vaden [00:15:54 – 00:16:20]

Um, what have you seen for, on, on the, the judicial side? And I imagine there’s probably a pretty big, pretty big split between, uh, judges and their personalities about adoption of, of new technology and wanting to do things the old-fashioned way. But for the judges that are like, let’s talk about the latest and greatest and what I can do to augment my job as a judge, what kinds of things are you seeing in that space?

Matt Davis [00:16:21 – 00:18:06]

Yeah, what we’ve seen in that space thus far is meant to assist the judge in their analysis and the summarization of their data. There’s a, there’s a company out there called Learned Hand, which is, which is interesting. They are, they’re doing the cross-referencing of the facts and of the references in the filings and raising attention for the judges. If, if, if they’re good, if they’re not, if they require more review. So I put this in the category of a, of analysis, right? This is not about outcomes. This is a, this is about analysis. Yeah. And if you zoom out on that a little bit, really what we’re looking at for the judges is, is what’s a way that you can bring them the right information that they need at the right time, at the level of detail that they need, so they can move faster, so they can absorb data more rapidly, so they can see risks and their, their challenges more So as an example, historically, a lot of judges, they write down all of their notes in the hearing, and they are doing that because they might need to write an order later, or they might need to remember where their mind was at when, when they next gather, right? But if you were able to get, say, the digital transcript, then you could use AI to do a lot of that automation, right? What’s, what’s the summary of what happened in the last round? They maybe don’t have to write as many notes because they can get it automatically out of the digital transcript.

Matt Davis [00:18:07 – 00:18:13]

So one of the things that we’re doing, we, we recently acquired FTR for the record, right?

Phillip Vaden [00:18:13 – 00:18:14]

Mm-hmm.

Matt Davis [00:18:14 – 00:18:35]

And that’s, that’s, that’s one of the major use cases. We, is that we want to make the lives of the judge easier every day. by bringing value from that digital transcript and surfacing information in a way that allows them to move a little faster and be more effective in their day-to-day work.

Phillip Vaden [00:18:37 – 00:19:28]

Yeah, that is, yeah. Congratulations on that acquisition. That’s a, that was a big one. And they have a really excellent product. The, yeah, I, I think that for the judges, I love the ability for them to do more robust analysis and on-demand analysis. I mean, that’s another thing that’s really nice is that as, you know, as sort of having an AI agent plugged in and going over your transcription of your meetings and things like that, that has shifted a little bit the way that I hold meetings and the way that I speak during those meetings occasionally from time to time, just to make sure we’ve got the point across on the transcript. I’ll make a note out loud. I’m like, so let me recap this very cleanly for the AI agent, what we want to do here.

Phillip Vaden [00:19:28 – 00:19:55]

And I imagine for judges to be able to just say, you know, okay, and then just like kind of couch that argument or couch the result of what both sides are saying and make a note, that is really, really handy for them to be able to have that same functionality that we have during our conversations, including this one. I mean, we could get a transcript of this and have it broken down into a summary in a very short period of time, um, after we’re done recording. So, um, yeah.

Matt Davis [00:19:55 – 00:20:48]

And it’s not only the summarization, but it’s, it’s the combining of data sources as well, right? If, if you’re a judge, maybe 15 years ago, you had a physical file folder on your desk with all the filings and reviewing that information and, and understanding it all and absorbing it all is a lot of work. Right? These days, those documents may be digitized, right? But you’re still looking at the history of the case and reading those documents, extracting relevant details, right? AI, and when you think about it, all that data could be in disparate systems that you as a judge use, right? AI allows you to bring multiple data sources into one location. aggregate information from all those, all those sources and bring a succinct view.

Phillip Vaden [00:20:49 – 00:21:22]

Yeah. Is, is any of that being used? And, and this, I guess this is part of the analysis piece is like, you know, could, can the judges say, hey, here’s the pleadings, here’s the, we’re gonna have a motion for summary judgment hearing. And both sides have, you know, filed their motions, filed their counter motions, filed their responses, replies, all that fun stuff. And then, uh, the judges basically plug that in with all of the attached, uh, uh, evidence exhibits and all that. And basically say, you know, find inconsistencies and, and, and peel those out. Is that some of the functionality that they have access to now?

Matt Davis [00:21:23 – 00:21:25]

That’s what Learned Hand does. One of the things—

Phillip Vaden [00:21:26 – 00:21:28]

That’s so good. Um, last touch.

Matt Davis [00:21:29 – 00:21:29]

Yeah.

Phillip Vaden [00:21:30 – 00:21:33]

The attorneys need to be, uh, real buttoned up now.

Matt Davis [00:21:33 – 00:21:34]

Yeah.

Phillip Vaden [00:21:34 – 00:21:38]

If you’re— ’cause I’d have real targeted questions for you if your stuff isn’t buttoned up.

Matt Davis [00:21:39 – 00:22:08]

And, and honestly, in, in this day and age, right, where the judge may receive AI-generated content, the ability to actually validate it a little more thoroughly is important, right? Because there can be bad references in there. There can be any, any number of things, especially with a record number of SRLs operating in the courts, right? It’s, it’s, it becomes hard to Make sure all those references and the cross-checks are good.

Phillip Vaden [00:22:08 – 00:22:22]

Yeah. Now, uh, for Learned Hand, for, for the people listening, a lot of ’em are, are, you know, they’re attorneys, they’re paralegals. Um, did, can they get access to this? Is this, is this available for them or is this a judges-only type product?

Matt Davis [00:22:23 – 00:22:41]

I don’t know for sure. Um, uh, I think it’s mostly aimed at the trial courts. Um, is, is their Their target area. Um, I wouldn’t be surprised if they’re also offering it towards, towards attorneys as well, because there’s, there’s a lot of value there as well, I would think. Yeah.

Phillip Vaden [00:22:42 – 00:23:22]

Yeah. ‘Cause yeah, I mean, if I’m, if I’m going into, uh, uh, to a hearing with a, a judge who’s using that product, I would certainly want to pay. I’d pay, gladly pay to be able to run it through there and see, is there something that’s, that I’ve, I’ve missed here that’s going to blow up in my face during that? during the hearing. Yeah. Awesome. So then we’ve talked about the clerks and we’ve talked about the judges. Are there any other personas in the courts and justice space or any other types of artificial intelligence adoption that would be kind of a nice to know or good to know, like things like Learned Hand for the practitioners?

Matt Davis [00:23:23 – 00:23:54]

Yeah, sure. Let me, I want to give one more use case that we see in the field right now and then maybe, you know, spend a little time on where we’re going here at Tyler Technologies and where we’re focusing our energy. So the other interesting use case is the Arizona Supreme Court. They’re using AI avatars to explain the opinions from the high court in plain language.

Phillip Vaden [00:23:55 – 00:23:55]

Really?

Matt Davis [00:23:55 – 00:24:09]

And so, and so They want, they want the public to understand these rulings. And you know as well as I that, that, that they’re very legal. They can be very hard to read for the general public.

Phillip Vaden [00:24:09 – 00:24:09]

Yeah.

Matt Davis [00:24:09 – 00:24:33]

So it’s, so if you go out to YouTube, you can find a bunch of AI avatars from, from the Supreme Court of Arizona explaining in plain language what the decisions of the court are. And I think that’s a really neat use case and one that is really beneficial and right in the heart of what AI is very good at.

Phillip Vaden [00:24:34 – 00:24:52]

Yeah, that is so good. Because there’s always, I feel like, you know, I’m sure somebody will start doing a similar thing for Supreme Court opinions, but that’s always, everyone’s very quick to become a constitutional a law scholar whenever any opinion is handed down.

Matt Davis [00:24:53 – 00:24:54]

Of course.

Phillip Vaden [00:24:54 – 00:25:20]

And they’ll take any, you know, people want to take their own extreme sides of any ruling. So having that kind of as close to straight from the horse’s mouth in plain language is great. Because yeah, when a court drops a 100-page opinion, the number of people that actually read it is probably 12, you know, in all of America. So having something that’s easily digestible by the public is really awesome. That is such a cool use case.

Matt Davis [00:25:21 – 00:26:01]

If you’ve never done this in the medical area either, I know this is off the subject a little bit, but when you get, get like a medical document with a bunch of really complex lingo in it, throw it into AI and say, explain this in plain language. It will, it is, it is really amazing what you what you get back. I think of those use cases as very similar where there’s a professional, you know, language that maybe the public doesn’t understand as much, but there could be a lot of value in them understanding it.

Phillip Vaden [00:26:01 – 00:26:58]

Yeah. Yeah. I mean, for personal health stuff, I mean, obviously I’m not using, I’m not, I haven’t substituted WebMD for Claude as my doctor, you know, but the, but yeah, I mean, it is incredible, you know, ’cause I mean, you know, health and taking care of This is important for every profession. And, as you get in things like you get your blood work panels in, and if you’re curious about beyond just is this in range or out of range, I’ve done things where I’ve loaded that in like, here’s my blood panel, let’s just pretend like it’s awesome. It’s usually not. But then, it’s like, hey, if anything that I’m a little low on or whatever, I’ll just say, what do I need to add to my diet? or my daily routine to help with these levels. And it’s pretty awesome. I then go and ask my doctor, does this make sense? And I’ve got a very good primary care physician.

Phillip Vaden [00:26:58 – 00:27:16]

He’s always very patient with me and my AI questions that I bring him. But it’s amazing all the things we have access to now that would’ve been science fiction you know, even 7 years ago. And so, yeah. Yeah. Awesome. Well, great.

Matt Davis [00:27:16 – 00:27:19]

Do you want to hear a little bit about what we’re doing here at, um, Absolutely.

Phillip Vaden [00:27:20 – 00:27:20]

Yes.

Matt Davis [00:27:20 – 00:29:00]

Yes. So we’ve got, we’ve got a good number of, of, of initiatives in this lane. We believe that AI can materially make the jobs and lives of the courts better at the end of the day. So when you, when you think about where we’re really focused, a lot of it is on automation and just efficiency for the staff with the objective being that they can free up their staff to work on more complex matters and really, really focus on the case outcomes and what their mission is rather than the tactical day-to-day of, you know, doing this task in a piece of software or executing this workflow. Right? So we’re focused on building things like agentic workflows for for very common tasks that the clerks may do, things that will just make their their day-to-day lives easier. It could be as you know simple as looking up the case balance or when the next hearing is on a case, or it could be as complex as automating the whole intake process for a new case, right? All of those things. But anything that’s repeatable, that there’s a rule set around, that requires less decision-making on the part of a human, those are the things that we are focused on. We’re also working on a new experience for the judges, right?

Phillip Vaden [00:29:01 – 00:29:01]

Okay.

Matt Davis [00:29:01 – 00:30:32]

With the acquisition of FTR, Being able to use the transcript along with the information from the case management system to bring the judges a unique and more comprehensive experience is, is, is really important. We think it can be a big game changer in their world. And then we’re doing some things in e-filing as well around automation of of that so that we can make it easier for the filers. And lastly, around the safety of the data, right? We talked about how the courts are the custodians of the data. It’s a big deal, right? They’ve got to ensure that it’s protected. In many cases, the protection of millions, millions of cases, right, relies on attribution or, or configuration that they do around those, those cases in their CMS, in their case management solution. And AI, it can spot if maybe one of those attributes is wrong, right? If there is maybe one of the juvenile cases that is marked public, but it shouldn’t be. Maybe that’s the way it is in the CMS.

Matt Davis [00:30:32 – 00:30:50]

It could be a, it could be a mistake, it could be an error, whatever it is. But if we can use AI to detect those things as early as we can, then we can, then we can help with the, with the confidentiality of the data for the courts and the public.

Phillip Vaden [00:30:50 – 00:31:40]

Yeah, that is great. Yeah, I know that, that, that was something that, that, uh, cuz I mean, it’s, it’s a, it’s a tricky thing with, with keeping up with that many millions of records. And you think about, you know, um, if, if one thing gets entered with the, the wrong case type or the security, um, uh, gets, gets modified on accident, you know, just a, a fat finger thing, then, you know, it’s, yeah, show, showing those, those cases. And that, that’s a, a big fear for, for a lot of courts. And that’s why some of them just, They’re like, hey, the safest thing is for me to, you know, not show these documents to the public. I can show the docket, but not the documents. So things like that, I think you’re gonna— sounds like that can lead to more open access to records because people have more confidence that, hey, we don’t have to worry about big mistakes that happened. You know, and sometimes it’s on the attorney side.

Phillip Vaden [00:31:41 – 00:32:04]

They won’t do the redaction work that they’re supposed to do before they file. And so being able to identify those things as, hey, there’s bank account numbers and Social Security numbers and, you know, driver’s license, you shouldn’t have this in here, you know? So obviously that document came in as public. Uh, all of that stuff should have been redacted, but we, we can just hide that document because it’s, it’s got this sensitive information. That’s really great.

Matt Davis [00:32:05 – 00:32:16]

Yep. We believe those are all really good use cases for AI that will, that will, um, make things better. In the court system. Yeah.

Phillip Vaden [00:32:16 – 00:32:40]

Awesome. Well, man, it sounds like you guys have a lot of exciting, exciting stuff you’re working on and, and already have available. Well, so one of the, I have, I have kind of a, a wrap-up question that I ask everyone, and that is, do you consider yourself a techno optimist or a techno pessimist? And what’s your, what’s your case for your position? Hmm.

Matt Davis [00:32:42 – 00:33:55]

That’s a good question. I would say I’m a little bit on the fence, honestly, right? When I was younger, I watched a lot of Star Trek. And in that world, right, humans are free of day-to-day toil for the most part because they have AI and they have, you know, all everything that they need to to meet the basics of survival, and they can then opt to spend their time where they want, where their passion is, right? And if you listen to Elon, he says that that day’s coming. It’s, he says it’s only 10 years away. I’m not sure it’s going to happen that fast. It would be lovely if it did, but he’s very optimistic, right? So I can see that path. I like that path. I would also love a world where I’ve got my own personal robot who can do my laundry and my dishes and make my meals and walk the dog, right? And if you think about it, I don’t think that is real science fiction right now.

Phillip Vaden [00:33:56 – 00:33:56]

Yes.

Matt Davis [00:33:57 – 00:34:39]

I mean, with Tesla building Optimus, which is their, their version of the humanoid robot that they’re going into mass production on, and the capabilities of the large language models that we see right now, I can, I can imagine a world where you put that large language model into one of those Optimuses and it can do all of those things. It can do your laundry, your dishes, it can make meals, it can walk your dog. I mean, that’s pretty cool, man. Yeah, that would be, that would be very cool. And I honestly think we’re going to see that in 10 years or less. Yeah, you know, I agree. And I agree. And it could be as little as 3 years even.

Matt Davis [00:34:39 – 00:34:41]

I don’t know, but I think we’re going to see that in our lifetimes.

Phillip Vaden [00:34:42 – 00:34:42]

Oh yeah.

Matt Davis [00:34:42 – 00:34:58]

Which is, which is going to be wild in my opinion. I think that’s real. Um, but we’ve all seen the movies, right? We’ve seen The Matrix. We’ve seen Terminator. We’ve seen WALL-E, right?

Phillip Vaden [00:34:58 – 00:34:58]

Yeah.

Matt Davis [00:34:58 – 00:36:18]

What happens when AI, when it gets a mind of its own and it decides that we’re the bad guy, right? And things happen. I would say right now I’m more of an optimist. And the reason I say that is because I’ve never seen AI have an opinion or an objective of its own. It is always executing on the objective that it was given by a human. Even if you think about the recent article from OpenAI where one of their models where it escaped from its containment and it hacked another website. Well, the reason it was doing that is because it was asked by humans to score the best it could score on the security benchmark for AI models. And it decided that the best way it could do that was to go, um, hack the results database and steal the answers because then it could score perfect, right? And so, and so it, it did some interesting things, right? It hacked out of its containment. It went and hacked into multiple websites.

Matt Davis [00:36:19 – 00:36:44]

But it was ultimately doing what we told it. And so I think we’ve got to get very good at saying what we want and understanding what the guardrails are. But unless it begins getting opinions of its own or objectives of its own, then right now I’d say I remain an optimist.

Phillip Vaden [00:36:44 – 00:36:45]

Yeah.

Matt Davis [00:36:45 – 00:36:45]

What about you, Phil?

Phillip Vaden [00:36:46 – 00:37:20]

Yeah, I’m an optimist. I know there’s going to be some negative things that come out of it, but I’m an optimist. I own a Tesla. I have full self-driving on it and I don’t really drive anymore. And this is a kid. If you would’ve told me I had a car that drove me around, I would’ve been like, I don’t know about that. And to my kids, it’s just like, it’s not a big deal. Like, this is just like, of course a robot drives you everywhere.

Phillip Vaden [00:37:21 – 00:38:06]

And so I look at it and go, you know, and so much safer. I then had to actually drive a different car the other day and it was extremely stressful. And I was like, I can’t believe humans are doing this. So it’s like, I don’t have eyes everywhere. I just have eyes facing the front here. So yeah, I think there’s so many use cases. I do think that the Optimist has a lot of amazing possibilities for making our lives, freeing up our time to spend more time with our families. Now the question is, and this is always like a personal thing to me, is like whenever you get this extra time, do you spend it on, do you spend it with your family or do you spend it doom scrolling on your phone? And so some of it, you know, it’s like personal responsibility.

Phillip Vaden [00:38:06 – 00:39:08]

Like what are you going to do with that time? And are you going to, you know, you mentioned the Star Trek example. are you going to chase those passions of yours? And I think for a lot of people, the answer is going to be yes. It’s going to take diligence, right? And so I think about for legal professionals, one of the things that I always wanted to do is when I was practicing the law years ago is I wanted to have more time for pro bono work because it’s like I’m representing large corporations and doing all this, people that can afford litigation. But I didn’t, I just didn’t have the time to go in and help out people who needed help but couldn’t afford it. And so I look at this and think, man, if attorneys all of a sudden can do 3x, 4x, 5x the work by being augmented by artificial intelligence, that will free up really good attorneys to be able to go and help people who can’t afford $800 an hour. And so I think it’s going to be a bright a bright future. Um, and, uh, it won’t be the Cylons, fingers crossed.

Matt Davis [00:39:10 – 00:39:11]

But will you buy an Optimus, Phillip?

Phillip Vaden [00:39:12 – 00:39:16]

Uh, oh, 100%. Yeah, as soon as it comes out. Yeah, I’m, I’m absolutely going to buy it.

Matt Davis [00:39:16 – 00:39:16]

Yeah.

Phillip Vaden [00:39:17 – 00:39:22]

And, uh, yeah, I’m like, I was like, even if out of the gate it’s just doing laundry or something, I—

Matt Davis [00:39:22 – 00:39:23]

that’s perfect.

Phillip Vaden [00:39:23 – 00:39:50]

I, you know, it’ll— because the other thing is like, you know, Tesla is very good about, uh, improving their products and just over-the-air software updates. So I’ve seen it with the car. And so now, you know, getting that robot and it’s like, hey, all it does is laundry right now. Perfect. Great. I know it’ll cook and put my groceries away and it can go to the store and do my grocery shopping and bring it back. I don’t have to do Instacart or whatever. It’s just, you know, one-stop shopping.

Phillip Vaden [00:39:52 – 00:40:10]

So anyway, well, Matt, thank you so much for the time, man. This has been really fun. I, uh, I’m probably gonna pester you about coming back on, you know, probably, you know, once or twice a year just to give us an update on, uh, on what’s going on in the, uh, the courts and justice space. So I appreciate your, your input and expertise, man.

Matt Davis [00:40:11 – 00:40:18]

Well, I appreciate the invite and I’m happy to be here. It was good conversation and I’d be happy to come back in the future.

Phillip Vaden [00:40:18 – 00:40:20]

Sounds good. Let’s grab lunch soon.

Matt Davis [00:40:21 – 00:40:21]

Okay, let’s do it.

Phillip Vaden [00:40:21 – 00:40:22]

All right.

Continue Reading