Quinten Steenhuis brings a builder’s mindset to legal innovation, rooted in early Indymedia activism where scavenged hardware became community infrastructure. That scrappy origin story carries through a dozen years of eviction defense at Greater Boston Legal Services, with a steady focus on tools that help people solve problems without waiting for a savior in a suit. Along the way, Quinten also lived the unglamorous side of mission tech, keeping systems funded, supported, and usable when budgets get tight and priorities get loud.
The conversation then jumps to Suffolk Law’s approach to generative AI education, including a required learning track for first-year students. Quinten frames the track as foundational training, then points to a deeper bench of follow-on courses and the LIT Lab clinic where students build with real tools, real partners, and real stakes. The throughline stays consistent, exposure alone solves nothing, so Suffolk puts reps, projects, and practice behind the syllabus.
A standout segment tackles the “vaporware semester” problem, where student-built prototypes fade out once finals end. The LIT Lab fights that decay by narrowing tool choices, standardizing around DocAssemble, and supervising work with a clinic-style model, staff stay close, quality stays high, and maintenance stays owned. Projects ship through CourtFormsOnline, with ongoing updates, volunteer support, and a commitment to keep public-facing legal help online for the long haul.
Then the episode turns toward agentic workflows, with examples from Quinten’s consulting work in Virginia and Oregon. One project uses voice-based intake to screen for eligibility, confirm location and income, gather the story in a person’s own words, and route matters into usable categories. Another project speeds bar referral by replacing slow human triage with faster classification and better user interaction patterns, fewer walls of typing, more guided choices, more yes-or-no steps, and fewer dead ends.
In the closing stretch, Quinten shares the sources feeding his learning loop, LinkedIn, Legal Services Corporation’s Innovations conference, the LSNTAP mailing list, podcasts, and Bob Ambrogi’s LawSites, plus the occasional spicy Reddit detour. The crystal ball lands on a thorny challenge for both academia and practice, training lawyers for judgment and verification when AI outputs land near-correct most of the time, then fail in the exact moment nobody expects. Quinten’s bottom line feels blunt and optimistic at once, safe workflows matter, and the public already uses general chat tools for legal help, so the legal system needs harm-reducing alternatives that work.
Listen on mobile platforms: Apple Podcasts | Spotify | YouTube
[Special Thanks to Legal Technology Hub for their sponsoring this episode.]
Email: geekinreviewpodcast@gmail.com
Music: Jerry David DeCicca
Links (as shared by Quinten):
Transcript
Greg Lambert (00:00)
Hi, I’m Greg Lambert with the Geek in Review and I’m here with our friend Nikki Shaver from Legal Technology Hub. Nikki, I know the maps that you guys create are a constant upgrade, but you just released a new one. Can you tell us a little bit more about that?
Nikki (00:18)
Yeah, that’s right. Hi, Greg. Hi, everyone. So we have been releasing our Gen.ai in Legal Tech maps.
every quarter since actually February last year, then we updated in March and then every quarter since then. So we released in early January, we released the latest one. The cutoff date for that was the end of December 2025. And it has 855 generative AI product placements by 682 vendors across 19 categories, which is obviously a massive landscape now. And in fact, back in March last year, I had said
a year from now we will have a thousand products on this map and I think we’re well on track for that, interestingly. But now that we have a year’s worth of these maps, it’s really interesting to see some of the trends and some of those are worth calling out. So first, the areas where we’re seeing the most startup activity.
That has changed over time. And in particular, this map shows a different type of activity than in the previous maps. So for most of last year, AI Legal Assistant was really the hottest category, a ton of concentrated startup activity, new startups coming to market in that category. In December, interestingly, that had cooled a little bit and the hot concentration of startup activity had shifted over into operational tools.
both law firms and legal departments. So for example, intake of both work and clients.
and also to compliance and governance solutions. This really tracks with what we’re seeing in the market in terms of where firms are focused. And it also suggests that maybe new entrants are seeing the broad assistant category as being relatively settled right now and also quite crowded. So that’s an interesting one to watch. The other thing I wanted to call out is we can really see, looking quarter to quarter over the past year, how the market is evolving.
The growth trajectory has been incredibly consistent and steady with about 100 companies added to the map every quarter. And so we include, you go and look on LegalTechHub, www.legaltechnologyhub.com, you have our clickable map, which is the full map that you can click into every logo, explore the categories. We also have a graph that shows that trajectory. And it’s just quite remarkable to see how steady it is, which is an indication
that we haven’t seen a slowdown in the market yet. Not at all. ⁓ No sign yet of a bubble bursting or anything like that. So we look forward to continuing to track that this year,
Greg Lambert (02:59)
All right, well good, my 401k is safe for at least one more year, right? All right, Where can they go to learn more about this?
Nikki (03:03)
Don’t take it from me.
They can go to Legal Tech Hub, so do visit our website, LegalTechnologyHub.com.
Greg Lambert (03:16)
Thanks.
Marlene Gebauer (03:24)
Welcome to The Geek in Review, the podcast focused on innovative and creative ideas in the legal industry. I’m Marlene Gabauer.
Greg Lambert (03:31)
and I’m Greg Lambert.
Today we are delighted to have with us Quinten Steenhuis who is the co-director of the Legal Innovation and Technology Lab at Suffolk Law School and also the owner of Lemma Legal on the show. ⁓ One of the things, Quinten and you can verify this, is that ⁓ in my research, one thing that stood out was it was basically saying that Quinten is a builder.
in a field that’s often filled with talkers. So we’re going to ask you to talk a little bit today though. So Quentin, welcome to the Geek in Review.
Quinten Steenhuis (04:04)
Okay.
That’s so nice to be here.
Greg Lambert (04:08)
So, Quinten your path to legal technology wasn’t through some big law innovation committee. started in the Indy Media Activist Collective in the early 2000s, and you spent like 12 years in the trenches of eviction defense at the Greater Boston Legal Services. So I’ve seen you framed as a
hacker activist ethos. And I think that helped kind of shape your approach in building the tools like you did, And then you stepped into academia. So how is it that this background with this, you know, kind of scrappy blue collar tech philosophy, how…
How did you enable that to then go into academia?
Quinten Steenhuis (04:57)
that’s an interesting take on it. And I appreciate that. So I’ll say my indie media days were during college and we were literally scrounging computer hardware out of the garbage, the dumpster, whatever the university was throwing away. Like, hey, there was tons being thrown away at the time. And we would take these and make servers for like an independent media collective in South America or Africa. So
Greg Lambert (05:10)
There was lots being thrown away at the time.
Quinten Steenhuis (05:22)
Yeah, there was some of that kind of that scrappy ethos for sure. I would say that what I learned doing that was really focusing on building tech tools that would empower folks to solve problems on their own. In indie media, that meant helping people report on themselves, talk about some of the activism they were doing in a way like kind of pre-social media.
without a filter.
Greg Lambert (05:46)
Yeah, this
was really early days. This was not like you could go on Facebook and get a lot of people. How did you do this and thrive as well as you did?
Quinten Steenhuis (05:59)
So we were doing this before Facebook, before YouTube, before there were really any good ways for you to publish your own photos online in the internet. kind of the Independent Media Center invented this idea of self-publishing, which has turned into blogging, turned into social media. It really took a very different turn from and expanded a lot beyond the original scope.
But it was just really exciting to be able to do this, to kind of say, hey, we can kind of do our DIY media, our own takes on things, let people kind of access the world unfiltered. And we just saw the positives at that point, and I’ll say we’re not responsible for all the negatives that kind of came out afterwards.
Greg Lambert (06:37)
Yeah, there was a good
time up until 2012.
Quinten Steenhuis (06:42)
pretty good run.
Pretty good run.
Greg Lambert (06:44)
How did you transition that into your activism eviction and housing industry?
Quinten Steenhuis (06:56)
So from that kind of tech, scrappy, DIY ethos, and as an undergrad, I decided to go to law school. thought, this is, really do like being able to think about how we’re changing people’s lives, how to impact the world. Technology did not seem like the right biggest lever for me at that time. thought law, this is a long history of reshaping the way people understand their rights and to move forward social causes.
So that pushed me towards law school, getting exposed to that and saying, hey, this is really great, but what more can we do? Maybe go to law school. In law school, that’s the first time I heard that there was this whole field called civil legal aid. I didn’t really know that it existed. I kind of knew about law from law and order, criminal defense, maybe some of the kind of small town law, right?
I was adopted, we had a family lawyer that helped with my adoption and that of some of my siblings. But that wasn’t really the same as that world of helping indigent people with legal needs. And I really connected with it right away, and I spent my time in law school in the legal aid clinics.
Greg Lambert (08:06)
And when you were at the Greater Boston Legal Services, you weren’t just practicing, you were also in charge of the technology from what I was reading. And one of the things that you talked about was your focus on making sure that the technology that GBLS had was good, that it was supported, that it was funded.
Quinten Steenhuis (08:16)
That’s right.
Greg Lambert (08:32)
And usually when you have organizations like that, those are some of the things that tend to get cut first or get money last. Do you think that gave you better success as far as learning to leverage the technology and helping the overall mission?
Quinten Steenhuis (08:49)
Absolutely. So one thing that does happen is that there’s lots of corporate philanthropy that’s directed at software because it’s essentially free for them to provide it to as many folks as they want. So we had access to some really big tools from Microsoft, sometimes from other partner technology organizations, Twilio, Amazon.
And that meant that as someone helping with the technology needs at my legal aid, wasn’t the only person, luckily, because I was practicing law at the same time. It would have been kind of scary if it was a solo tech operation. I was able to experiment with these big tools and see what was out there. It’s kind of interesting, different perspective, We were low resourced, but we’d access to these really big enterprise level tools for free or very low cost.
Greg Lambert (09:34)
It’s interesting.
Well, it’s interesting because it’s kind of like a free puppy, right? Yeah, you can get it. But if you don’t have the expertise and knowledge and time to invest in the technology, it sits on the shelf and does nothing. I’m sure you and whoever else was helping you with the technology brought great benefit to that as well.
Quinten Steenhuis (09:40)
Right.
Marlene Gebauer (09:41)
take care of it.
Quinten Steenhuis (09:57)
Yes, that was definitely something to figure out. Like, hey, I can experiment with this and figure out is this too big for our needs or is it something we can scale? It was a big legal aid program. there are more than 200 staff and attorneys there now. It was a bit smaller at different times. The reason why I took on that second tech role was because of the financial crisis. We’re probably at our smallest in those years that I was also doing tech work. But.
Yeah, we could use these tools, but we had to be careful for sure to know that they weren’t too big for our size of organization.
Marlene Gebauer (10:31)
So I want to fast forward Quinten to present day and of course talk about Gen AI in the law because we have to do that. ⁓ Suffolk Law recently announced that all first year students will complete a required generative AI learning track in their legal practice skills course. From your vantage in the lit lab, how are you ensuring this isn’t just
Greg Lambert (10:40)
required.
Marlene Gebauer (10:53)
you know, check the box exercise, but it really equips students to use AI tools thoughtfully, of what elements of the learning track or follow-up focused on and proud of?
Quinten Steenhuis (11:06)
One of the things that I really like about working at Suffolk is how well integrated some of these different technology electives are into the core curriculum. So I do really like this class and I think it’s fair to say, hey, you know, this is kind of like a first year writing class. We’re learning the basics, we’re giving everyone a foundation to build on. So for some folks, it might be a little bit of a checkbox exercise to learn just those foundational skills. But then if that student gets exposed to something,
They know how to use those technologies throughout the rest of their time at Suffolk. And they have many courses that they can build on to take that forward and further. So we have an AI in the law class that kind of dives into some of the policy questions and issues with the hands-on element. I myself teach a seminar, which is legal technology for small firms, where we’re using about half of the course now today since 2023.
been adding more and more of this generative AI content. And now it’s about half of the thing that we spend our times on. We still hit the other fundamental pieces, but about half of the time probably is on generative AI. And then in our lab, we take students and give them a full year clinical experience, 12 students that are learning and building with these same tools. So I think that’s what we do is we’re taking it from you get that exposure, you understand the fundamentals and the basics.
And then there’s a whole three years of content and learning that you can actually put these tools into practice. That’s what we do. We’re all about, as you said, kind of building and doing things is kind of core to our philosophy at Suffolk.
Greg Lambert (12:44)
And you’re partnering with Hot Shot on some of the training, is that right, with Ian Nelson? ⁓ How is that partnership going? Are you able to contribute back so that they can advance that? I’m curious what the cycle is for that.
Quinten Steenhuis (12:51)
That’s right. Yeah.
That is a really good question and my colleague, Diane O’Leary, probably be a better person to answer it. But from what I understand from talking with her about it, because she runs the Legal Innovation Technology Center, which kind of is in charge of some of those core classes. But my understanding is that there’s going to be some mutual benefit there for sure.
Greg Lambert (13:17)
Good one.
Yeah,
maybe we’ll get Diane and Ian on the show together and we’ll talk about that. That’d be fun. You know, one of the things I think a lot of us that have gone through any of the…
Marlene Gebauer (13:22)
Yeah.
Quinten Steenhuis (13:24)
Yeah, that would be great.
Greg Lambert (13:35)
innovation labs or even.
practices at law schools where it’s a one semester, maybe two semester thing is that you get this almost vaporware problem going on where you might have a student that comes up with a really cool idea, that they develop something and they prototype it, and then the semester ends and then the tool dies. They go on, they get the job somewhere else and attention changes.
Yet somehow in the lit lab, you’re maintaining these tools like the housing search log and court forms online. So how is it that you’re structuring the lab to not just create the innovation and the idea generation, but rather sustaining that over a while? Do you mind?
giving us like an example, I saw where Anna Morissette did ⁓ this tool on housing vouchers. I mean, there things that, how do you do the process and then what are some things that are ongoing?
Quinten Steenhuis (14:46)
Absolutely. So when I joined the lab, I think we had kind of a range of different focuses and different kinds of tools that we were using for projects. And one of the things that we did is that we decided to streamline our focus and to really zero in on a couple of core technologies that we really felt, yeah, we can support this. It’s not something that’s going to take a lot of relearning to see, well, what did that person do again? How did it work?
if we need to make updates to it. So we focus on using DocAssemble. DocAssemble has just been in its 10th year now, which is kind of amazing, built by ⁓ basically a solo software developer who’s a legal aid attorney in Philadelphia, Jonathan Pyle.
Greg Lambert (15:28)
Well,
before you go on, talk about that because that’s really interesting. Talk about the dock assemble and how that’s still going on, what it is.
Quinten Steenhuis (15:31)
Yeah.
Marlene Gebauer (15:32)
Yeah.
Quinten Steenhuis (15:37)
Yeah, so I started using it as a legal aid attorney at Greater Boston Legal Services. And at that point, I think it was like just two years old. And I decided, hey, this is the really neat software. makes it easy for you to build user-facing software tools, like TurboTax for X. So I built a tool called Made that was, in a way, TurboTax for eviction defense, helping attorneys maybe with a client or…
other kinds of non-attorney advocates, as well as tenants directly to help fill in all of the eviction defense paperwork that they needed to do accurately, correctly. It’s really nice, easy to use software on the front end. It’s a little bit like it looked 10 years ago, but it had a really advanced enough user interface that it still looks good today, and it’s very flexible and customizable. So, yeah.
Greg Lambert (16:27)
Yeah, it’s open. It’s open
source and it’s on GitHub, right?
Quinten Steenhuis (16:30)
That’s exactly right. Yeah. So anybody can run it for free. There’ve been several commercial software products built on top of it because it’s available for free. Gavel.
The other big one at the time that they were big participants in the community is now owned by NetApp.
community lawyers, what they started out as, and they went through a couple of renames. But I’ve heard from a lot of other smaller tech vendors that they’re kind of quietly using DocAssemble under the hood too, because it does have this free and very flexible open source approach to doing it.
Greg Lambert (17:03)
That’s why it’s there, that’s why it was built.
So do you have some examples ⁓ of things that your students have built that you’re continuing to maintain?
Quinten Steenhuis (17:13)
And I think that’s another thing that’s kind of key to the way that we do things is so students will at the beginning of the year, they get a small project that we say, all right, this is going to be launched this semester. And our students have just turned those in for this year’s clinic. And they also have a larger project that they’re integrated in where they pair up and they need to finish it by the end of the year. And they work with court partners to do that. But as clinicians myself,
David Colaruso, Sam Glover, Bryce Willey, we all help make sure that that software is delivered in a scalable, maintainable way. Just in the true clinic fashion, right? You don’t just send your student out into the world to represent a death row client all by themselves. You’re there with them every step of the way of making sure everything is done correctly. And we do that from both the relationship building and maintenance side, as well as the software building side. And all of those tools go onto a website that we built during the pandemic.
It’s called courtformsonline.org. These tools get regular updates and review. We have volunteers who help with some of those steps and we take that on ourselves and just guarantee once we launch it, we’re committed to keeping it going indefinitely. And because they’re all kind of built around this core set of tools and frameworks, we actually feel like we can continue to make that commitment on an ongoing basis. We kind of pick tools we know we can work with that they’re gonna stay around.
10 years is a pretty good track record already. I think we can probably stay around for another 10 years with these tools, maybe 20 or 30. And we help make sure that everything is kind of done in a way that can work.
Greg Lambert (18:48)
I’m definitely
seeing the doer part in what you’re doing. So good job.
Marlene Gebauer (18:52)
⁓ And I’m
Quinten Steenhuis (18:52)
Yeah.
Marlene Gebauer (18:53)
curious, so when students sort of go into the program, how do you see them sort of change by the time they get out, and what sort of feedback are they giving about the experience?
Quinten Steenhuis (19:06)
So students really value the clinical experience at Suffolk. And I will say that sometimes they don’t really know a lot about what a clinic will be. They know they want to be part of a clinic. They might not necessarily self-identify as a very technical person. I’ve seen some of our folks who entered as the least technical, least self-identified as being technical be our best performers in the clinic. So it’s been a huge transformation for some students. Sometimes…
They start by taking a seminar like mine, and then they realize, I really like this stuff. Let me keep going. I’ll take the clinic. Or they might even have taken the clinic and think, I should probably learn more about this, but not really be sure or confident. They get to know, they don’t have any prereqs, so they can join and they learn from the beginning start to finish how to build software tools with our support. And it can be really exciting for a lot of them to be able to see that output of their work.
Greg Lambert (20:00)
Are you finding, because I see it here within the law firm, is that the trait isn’t necessarily the most intelligent person, but rather it’s the one that’s willing to be very curious, to try things out. Are you finding that to be the trait that leads to success in your clinic, or is it something else?
Quinten Steenhuis (20:21)
Yeah,
I would say so. think that curiosity is really important. There’s also this thing that we look for that I didn’t realize, you know, not everybody has this kind of ability or interest or willingness to say like, this is a big goal I have, I’m going to keep going towards that goal and figure out how to get there. So that’s something we really encourage is folks. Perseverance. We talk a lot about just having
Marlene Gebauer (20:43)
Like perseverance.
Quinten Steenhuis (20:50)
kind of a growth mindset for what they’re doing, but also just knowing, hey, I can’t just wait for someone to tell me the next step. I have to learn it and figure it out. We’re going to support you. We’re going to know how to help you get there. But you need to do some of that generation of like, these are the things that need to do be done to get here. That might be the most transferable skill that we can give our students is helping them do that growth. And we try to do in a supportive way. don’t want to just, we don’t leave people out to do everything on their own.
We want to encourage their ability to do things on their own.
Marlene Gebauer (21:22)
So I want to talk a little bit about, um, the Suffolk innovation community. Um, and you know, you’re deeply, obviously deeply embedded in this community. So, you know, you are a judge in the American legal technology awards with Tom Martin. Um, you collaborate with colleagues like Diana Leary, who you mentioned and, and Gabe Tenenbaum, um, you know,
There’s often a tension between sort of the slow, deliberate pace of legal academia and like, I’m sorry, my bias is showing here a little bit. You can correct me if I’m wrong. ⁓ and the move, you know, and sort of move fast, make mistakes, you know, move on, culture of, of how does the ecosystem at Suffolk manage that type of tension? If there is that tension.
Greg Lambert (21:52)
Hahaha!
Quinten Steenhuis (21:53)
Thank
Greg Lambert (22:08)
How do you keep
from punching somebody? Is what she’s really saying.
Marlene Gebauer (22:10)
Hehehehe
Quinten Steenhuis (22:11)
You
Marlene Gebauer (22:12)
Asking for a friend, right Greg?
Greg Lambert (22:14)
Hahaha
Quinten Steenhuis (22:15)
We’re lucky to have a dean who’s very supportive and interested and kind of leads the faculty, which I think could be the opposite at some law schools, probably. But Andy Perlman, he really is a national leader, international leader on issues of law and technology and really is very supportive. So something that I’ve also made a very conscious effort to do is to help break down coming from a very doer oriented.
world, the legal aid world, where people are trying to figure out how do we do with a little bit of resources, help as many people as we can. I tried to bridge the gap between that and the academic world. We’ve been doing these regular AI and access to justice workshops. I actually just came back Saturday from Italy where we had the fifth edition of our AI and access to justice workshop at the Eurex conference in Turin, Italy. And we presented some of the work that I’ve done there.
And what’s happened is that we get a mix of folks participating in those workshops, folks that are, it’s mostly kind of traditional papers that people are presenting about work that they’ve done or ideas that they have. But we have a mix of practitioners and folks that are in this traditional academic circle. And it’s really nice cross-pollination that happens there. I think that we have folks that are really yearning for both, but you have to kind of break through the normal way that people
interact, which is through these social circles, and figure out how do we get them to talk to each other in this kind of setting. I think works pretty well. There’s other settings. Margaret Hagen, who’s at Stanford, she helps run these workshops too. We’re doing with also folks at Carnegie Mellon, the Maastricht Law and Tech Lab in the Netherlands, kind of form this core group of folks running these particular workshops. But Margaret, I think, had a workshop.
That was mostly the legal aid community just a few weeks ago. And this community here is mostly academics with legal aid folks and other practitioners brought in as well. I think that’s the way to do it, because people hunger for, the practitioner side, they want to understand, well, what’s the leading edge research? What are the ways I ensure the quality and safety of this? How do I understand, does this thing work? People launch things without necessarily knowing that all the time.
And then on the academic side, they’re really hungry to say, like, why is the work that I’m doing relevant? How do I actually know if this thing is usable? Do people actually know which buttons to click? And they can see, they can make choices that kind of will lead to them actually using it, actually having it launched and in production. Both of those sides are really interested in hearing from each other. And it makes sense that we spend most of our time talking to each other because these are very, they’re strong skills to build and to.
that you can spend all of your life on just one side of this problem. But they want to hear from each other. So that’s one of the things I think we’ve really been successful at Suffolk is helping foster that community. And I think it’s really critical to kind of have that kind of ability to talk to each other about what we’re doing.
Marlene Gebauer (25:16)
The first step is to bring everybody together and see where the synergies go.
Quinten Steenhuis (25:22)
Absolutely.
Greg Lambert (25:23)
Do you hear that law firms and clients? So, know, Clinton, over the past few years, past three years, in fact, we’ve been really focused on the AI chat bots, the, you know, giving it something, getting something back and then going back and forth.
Marlene Gebauer (25:29)
content.
Greg Lambert (25:44)
⁓ and you know, it’s like asking a chat bot to go write a legal memo. ⁓ but a lot of us, including you are now talking about agentic AI where the AI tool can actually independently go out and take some action. so where do you see a shift to where we can get these AI agents to begin to autonomously handle, you know,
administrative burdens of legal aid, say like intake or scheduling or eligibility checks. Is the future of legal tech less about the requesting for a memo to be drafted and more about go do this and finish this task?
Quinten Steenhuis (26:27)
Absolutely. I think agentic AI is like a really interesting term because it sounds really fancy, but as you said, right, it’s really just AI that can do things for us besides just like responding with text. So I’m working on some of these tools with my consulting hat. So I’m working with Virginia Legal Aid Society and the Oregon State Bar. That’s actually the work that I presented last week.
in Turin and actually won an award, which was kind of nice. The paper won an award from that conference. So what we’re doing for Virginia is that we’ve built a AI voice-based intake system. And it goes from start to finish of the screening steps of helping somebody get matched with an attorney. So it can help them confirm their location as one that the
Marlene Gebauer (27:00)
Congratulations.
Greg Lambert (27:01)
Congratulations.
Quinten Steenhuis (27:24)
Legal Aid Society can help with. It confirms things like their income. And it also gets them to use their own words to describe their legal problem. You can ask a couple of follow-up questions to kind of zero in on what’s really happening with this person’s case. And then it uses a tool that we built to help classify that person’s legal problem from a long list of possible categories. So some of those might be things that the Legal Aid Program helps with and some may not be.
Sometimes they might be talking about something happening in their life where there’s not actually a lawyer that can help them. I talked about this. It became a particular issue when we were helping the Oregon State Bar. So, you know, and you’ve probably seen this. If you have an email address on the internet, you might have gotten a really out of the blue message from somebody who really wants to talk to you, but doesn’t have something going on where they can actually use a lawyer’s help. They might be dealing with mental illness, whatever it might be, right?
We have to handle those really sensitively, but we want to help screen those people out as well. Yeah. So we then they get to talk to an attorney at the end.
Greg Lambert (28:23)
sense. Yeah, needed that when I was at the Oklahoma Supreme Court.
We would get all kinds of interesting requests on things that
Quinten Steenhuis (28:32)
yeah. Yeah.
Greg Lambert (28:32)
Yeah.
Marlene Gebauer (28:33)
But that was in person, right? Or was that online? That was email. All right.
Quinten Steenhuis (28:39)
The most persistent people in the world are the people who have the problems where everyone’s already told them no, I guess, right? And you want to figure out how to help them, but often it may not be the right resource for them. Yeah. They’ll find a way to contact you if you have a name or an email somewhere that they can find.
Oh, well, the other half of that project, I can say, it was just for the Oregon State Bar. We’re helping with their bar referral service. And it’s using that same technology, that legal problem matching technology. So we help them use their own words. We’re not using voice for that project. It didn’t really seem like the appropriate technology to use, but it gives them an instant referral. Whereas what would happen before is that someone would put something in in a web form, attorney would look it over, and then take a day or two to give them a referral.
instead. So really cutting out that middleware. And to that idea of like, agentic AI, if we’re talking about chatbots, chatbots use this idea of like, you have to kind of generate a bunch of words and respond back with those words. With these new tools, we could give people a form to interact with instead. And it can be generated by AI. I think that’s a really powerful, different interface that can help people.
be more efficient when they’re using AI through a phone, for example, instead of being asked to type a bunch of words, you listen to what they have to say, you respond with, okay, well, these are three choices that I think match what you’re talking about. Can you help me choose which is the right one that you mean? Ask some yes or no questions that again, the person could just click rather than having to type out an answer. And so we’re really playing around with these ideas to improve the user experience of interacting with AI instead of.
being asked to chat or type a lot, making that a more pleasant experience. And I think these are really already being used today. We’re using them today. And I can just see new metaphors and ways of interacting with AI that help make it a better experience for everybody coming in the future.
Greg Lambert (30:33)
On the voice, what particular software are you using for voice and are you finding it to be pretty effective?
Quinten Steenhuis (30:42)
That’s a really good question. I, Docassemble has a cat as its icon. So of course the next thing to do was is there another open source platform we can build on that has a cat as its logo? And there’s a tool called Pipecat. It’s built by a commercial company, but this is kind of the foundational technology that they don’t, they don’t need this platform to sell licenses to make money. So they release it for free.
They have other services they build on top of it, but lots of people like us can benefit from it. It’s updated very regularly and it’s very modular. So we, as new AI models are released, we can experiment with embedding those into the tool. And it uses this loop from speech to text. We have the AI do its piece. Then there’s text to speech.
So we’re really able to use the best for each of those different parts of the process. I think it has a lot of promise. Actually, at the workshop last week, one of the other projects that was discussed was a voice powered AI agent for folks in Niger. And in that case, these are very small language populations that might be using the tool. There’s low literacy. There’s very low access to technology.
so folks don’t have internet access. They don’t have a smartphone. Really the most advanced technology they have is a phone. I hadn’t really thought about how key that was as an enabling technology, the fact that now we have AI that can just use your voice to talk to it and respond back to with information as far as bridging that digital access gap. So I think that’s exciting as well.
Marlene Gebauer (32:19)
So before we get to our crystal ball question, we, and I mean, I’m sure you have lots of resources given our discussion. You know, what do you rely on to learn and to keep up with the rapid changes in legal tech, AI and the business of law?
Greg Lambert (32:36)
Yeah, where do you find all these cool free toys to play with? ⁓
Marlene Gebauer (32:38)
free toys and, and, and doing ideas, you know,
other than going to Turin
Quinten Steenhuis (32:44)
So yeah, great question. really like, I find myself relying on a lot of pretty old fashioned ways to learn about new things. So I used to be really active on Twitter. That platform I think is pretty for me. I’m done using it. I think a lot of other folks are too. A lot of our community has moved to LinkedIn. I really like to follow and see what people are doing there.
And that’s a key way I stay in touch with people other than at conferences, which are a couple of times a year. I really love the upcoming innovations and technology conference that the Legal Services Corporation puts on. It’s always a great way for me to learn about new things. In between those kind of capstone events, I also use some mailing lists like the Legal Services National Technology Assistance Project mailing list. I love podcasts like your own.
It’s a great way to learn about things and say Bob Ambrogi’s Lawsites Blog is another key resource site I go to. So I try not to procrastinate too much by reading all this news from the fire hose every day, every minute, but I do appreciate it. ⁓ I would say Reddit is in there too. A lot less frequent. It’s a much smaller community for legal tech on Reddit and sometimes.
Marlene Gebauer (33:56)
Sometimes they get a little spicy. You do.
Quinten Steenhuis (33:56)
lower quality and yeah, you learn about some things. Sometimes you see things there
on the other bigger platforms a little bit later.
Greg Lambert (34:03)
There’s all kinds of rabbit holes you can go dive into. All right, well, we’ve come to the crystal ball question, Quentin. so pulling out your crystal ball, looking out into the near future, do you think is the biggest change or challenge on the horizon?
Quinten Steenhuis (34:05)
Absolutely.
Greg Lambert (34:24)
that the legal industry or maybe even legal academic ecosystem itself needs to be prepared to handle.
Quinten Steenhuis (34:33)
In academia, I think our big challenge is going to be how do we help people learn the things that they need to learn when there’s an AI that can do those things for them very quickly, but just not perfectly. And they need to check the work. So it’s really hard to figure out how much time we spend on practicing rote skills that the AI can do really well 90 % of the time and help them to be prepared to catch the 10%. I think there ways, I think we’re going to see new learning strategies that
really take advantage of how AI can help us. It can help with simulations, practice sessions, giving people instant feedback in a way that one law professor couldn’t possibly keep up with to give really tailored education. So, but it’s going to still be a challenge. And then on the industry side, I think we have to really confront the same thing, that AI is not perfect at the things that we ask it to do. So we hear a lot about hallucinations.
I think we’re going to see more and more strategies focused just on that. But they’re kind of core to the way that large language models work. We need to add something as an extra check to help solve that. Other quality issues too, not just hallucinations. I think that engineering around that is going to help us advance the capabilities of these tools. For me, I focus on safe use cases. What can we do that’s reliable, testable, end to end?
But that doesn’t escape the fact that self-represented litigants are just going to ChatGPT and saying, write me a motion or help me file this claim. And the AI is just nodding along and saying, yeah, of course, that’s what you need to do. And generating sometimes nonsense documents that the court has to respond to. We can’t just say, hey, just use AI in these safe ways and then don’t give people an alternative. So I’m curious what an AI maximalist perspective will get us and how we get there.
Greg Lambert (36:08)
Thank
Quinten Steenhuis (36:22)
And how do we kind of do a harm reduction against the alternative, which is really just people going wild with ChatGPT and doing whatever they want. So we need to figure it out, helping people do it safely.
Marlene Gebauer (36:34)
Well, I think the future is bright with folks like you thinking about where we’re going to go. So, ⁓ Quinten Steenhuis thank you so much for taking the time to speak with us.
Quinten Steenhuis (36:44)
Thank you so much for having me on. I really enjoyed it.
Marlene Gebauer (36:47)
And of course, thanks to all of you, our listeners, for taking the time to listen to the Geek in Review podcast. If you enjoy the show, share it with a colleague. We’d love to hear from you, so reach out to us on LinkedIn and Blue Sky.
Greg Lambert (36:59)
And Quentin, is there any particular place that you’d want to point listeners to to learn more about what you’re doing there at Suffolk or if you’re your company or you specifically?
Quinten Steenhuis (37:10)
Sure, yeah, I can just give you two links. So if you go to the Suffolk Lit Lab L-I-T lab.org, you can learn about Suffolk’s projects in this area. And then you can reach my consulting business at Lemma Legal, L-E-M-M-A, legal.com.
Marlene Gebauer (37:28)
And as always, the music you hear is from Jerry David DeCicca Thank you, Jerry.
Greg Lambert (37:32)
All right, thanks, Jerry. Thanks, everyone.
Marlene Gebauer (37:34)
Thanks everyone, bye.