Dr. Megan Ma On The Future of AI in Legal
Dr. Megan Ma On The Future of AI in Legal | S1:E2
Dr. Ma shares insights into how generative AI is transforming the legal sector, from automating contract drafting to developing AI-powered legal personas that simulate partner-level expertise. She highlights the growing fascination with legal technology, driven by AI’s ability to process vast amounts of text and capture the nuanced behaviors of legal professionals.They also dive into the impact of AI on legal education and training. Dr. Ma discusses how law firms are rethinking mentorship in an era where AI can handle much of the routine work traditionally assigned to junior associates. She introduces groundbreaking tools like negotiation simulators that allow lawyers to refine their skills by engaging with AI-driven legal agents.Omar and Dr. Ma further discuss the challenges of AI adoption in law, including concerns over hallucinations, accuracy, and trust. They examine the shift from viewing AI as a back-office tool to a fully integrated system that enhances legal decision-making. They explore the changing role of Chief Legal Officers, who are increasingly taking the lead in AI-driven transformation. As the discussion unfolds, Dr. Ma offers a bold vision for the next decade of legal tech, where AI seamlessly integrates into everyday legal practice, reshaping workflows, client interactions, and even the very definition of what it means to be a lawyer.
00:00:00:02 - 00:00:23:05
Omar Haroun
Welcome to the Augmented Intelligence Podcast. I'm your host Omar, here in the co-founder and CEO of Eudia. I'm really, really excited today to welcome Doctor Megan Ma from Stanford. Megan is a pioneer in the legal technology space and is doing really cutting edge research on AI agents, augmented intelligence. A lot of things we're going to get into today.
00:00:23:05 - 00:00:43:10
Omar Haroun
And just as a very deep thinker on how lawyers actually kind of operate. Having had a lot of experience now with a PhD in law and and coming at it from a very academic angle, but still being very close to some of the realities of, the work that Codex and Stanford is doing. And so, yeah, I'm really excited to welcome Megan.
00:00:43:12 - 00:01:04:23
Omar Haroun
Megan, welcome to the show. I mean, yeah, of course, I, would love to start with with kind of what we were just joking about because both of us had makeup being applied to us. We're in this fancy studio, and it suddenly feels a little bit like, we're we're people that other people actually want to listen to.
00:01:05:00 - 00:01:19:06
Omar Haroun
Which is not the case for the last, you know, kind of a decade, I would say. So we would love to hear why you think there's been this sort of sudden explosion of these many people who are suddenly interested in, in legal tech for the first time.
00:01:19:08 - 00:01:45:17
Dr. Megan Ma
Yeah, that's a great question. And I, I always think that, you know, when I started, and even in the days of my PhD, my supervisor XR was the only one that believed in me that thought there was any value in the work that I was doing. I think I had pitched to our committee so many times, like, why there is going to be automation in the legal industry and why there's going to be this role that technology is going to play.
00:01:45:18 - 00:02:12:10
Dr. Megan Ma
Instead they were like, well, you know, nothing you're saying is about the law. And so I think today, like, they actually showcase that part of the work that the institution is doing is technology related. But I think kind of what makes people fascinated about the legal domain all of a sudden is with generative AI. They felt like, oh, you know, what exactly do lawyers really do?
00:02:12:11 - 00:02:49:06
Dr. Megan Ma
There's kind of this opacity in this smoke and mirrors, what? The whole profession seems to be tethered around language. And what makes kind of language models operate? Well, you know, if you process amounts of text, like, maybe if I generate things and can speak in legalese, that's all that there is. And so I hear, like a lot of young founders today being like, oh, well, there's all these inefficiencies in the legal system not knowing anything about kind of like the legal background, but simply that if you just throw it all into a model, then they're going to be able to mimic it in perfection.
00:02:49:08 - 00:03:22:22
Dr. Megan Ma
And I think like slowly, you'll see that all of the questions around, why this technology is actually not only fitting, but at times kind of, sufficiently nuanced for our profession actually is not founded in purely the data itself. It's actually understanding the lawyers and how they behave. And for the first time, we have a technology that's capable of capturing behavior as opposed to purely about the text alone.
00:03:23:00 - 00:03:53:11
Dr. Megan Ma
We have like, founders that are like 18, 19 that come to our existence and they have no legal background, but they're super excited about the legal domain, simply because they can speak in legalese. And it's kind of funny because when I started teaching, I taught a class called, You know, How to Speak the Law and the kind of question that a lot of students came back to me was, oh, I'm only taking this class because I feel like I want to integrate legalese into my day to day.
00:03:53:17 - 00:04:13:17
Dr. Megan Ma
And I was like, what does that even mean? They're like, well, it makes me sound unapproachable. And I thought that that was kind of like, a strange point to come across. And I found that I think like that continues, with a lot of engineers where they're just thinking that we're speaking almost like a different dialect.
00:04:13:19 - 00:04:32:12
Dr. Megan Ma
And in many ways it is. But, once they actually get into selling the technology or thinking about building in the profession, they realize that so much of it is actually about the characters of these lawyers. And how they actually present the knowledge to the world.
00:04:32:14 - 00:04:45:23
Omar Haroun
Fascinating. And and speaking of that, I think some of the research that you've done more recently with law firms where you've actually gone almost on the individual partner level. I think it'd be fascinating for people to hear about that.
00:04:46:01 - 00:05:06:21
Dr. Megan Ma
Yeah. So I think one thing we have been exploring a lot on is this notion of generative agents. There's been this idea that we're kind of classifying as legal personas, and it's this idea that in the law firm, it's not just about the law firm level of language or the law firms precedent. It's about at the partner level.
00:05:06:21 - 00:05:28:10
Dr. Megan Ma
Right. So you enter into a law firm and you're kind of green in the field and you're not sure what to do. It's very rare that you immediately start tackling like from just generic templates. Instead, you're kind of following the path of a specific partner. And when you go into these one on ones with partners, partners will kind of get back to you and say like, no, you've done it all wrong.
00:05:28:11 - 00:05:55:00
Dr. Megan Ma
They're partners that say, I don't like to use the word vary or whatever. And over time you kind of pick up the habits of the partner. And that's what we started to kind of touch upon, is in a world where a lot of this technology is going to take away the core component, that we used time and time again to teach our young lawyers, which is learning by doing, telling them to do thousands of contracts.
00:05:55:00 - 00:06:15:00
Dr. Megan Ma
And then at some point it will unlock, than being a senior lawyer if all of that goes away, because our tools are now available to generate contracts or do first drafts or whatnot. What are you using to teach the lawyer? And you can't just be teaching at a generic level. What you're what you have to teach is an apprenticeship model.
00:06:15:02 - 00:06:39:10
Dr. Megan Ma
And so a lot of the tools that we started building is how do you download the brains of the partners, capture their war stories, and actually simulate those experiences into the actual legal workflows that they do? So we built, for example, a negotiation simulator where you can actually simulate what it's like to negotiate an M&A deal. You can play on different modules, you can play on different modes.
00:06:39:10 - 00:07:02:19
Dr. Megan Ma
And what we found was really fascinating is that, partners started to get interested in it because they, they were actually negotiating with themselves and gaming their own strategies. For a junior, they started to see different like levels of aggression. So they saw, for example, if you play on Mistral model, they noticed that they almost always walk away from the deal.
00:07:02:19 - 00:07:22:20
Dr. Megan Ma
It could be related to the French training data or the kind of French character of it. But, you know, all of these things they started to pick up are what people broadly defined as soft skills, but actually are really the value or relative value differences. That's beyond, again, the four corners of the contract or four corners of any legal tax.
00:07:22:22 - 00:07:43:07
Omar Haroun
Fascinating. And so I mean just just to kind of pick up on the evolution of the industry, one change is there's now a lot of people, engineers in particular who see this is a lot of tax. It's a great candidate to use language models. And I get to speak any language. What are some other trends that you've seen in the legal technology industry?
00:07:43:09 - 00:08:10:09
Dr. Megan Ma
Yeah, I think in the past we have seen like legal focused on expert systems, that there was a huge kind of motivation for absolute precision and that you had go through a number of dependency trees. Of course, the struggle that they worked with was, scalability. You can only do very, very niche areas. The kind of other spectrum was a movement towards standardization.
00:08:10:14 - 00:08:34:00
Dr. Megan Ma
They were like, oh, we can just build contract playbooks, for example. And but that required kind of something that was to me much harder, which is being able to articulate some of that implicit knowledge that you as a senior expert, can't quite define clearly. And so how do you kind of describe top down a process that you almost are automating in your mind?
00:08:34:02 - 00:08:55:00
Dr. Megan Ma
I think this current era and why it's fascinating is that it picks up on that implicit knowledge very quickly. And so you don't need a ton of examples, in say, like older types of NLP and using NLP in legal tech. There were a lot of issues with kind of consistency of extraction, and information retrieval.
00:08:55:02 - 00:09:21:19
Dr. Megan Ma
And obviously a lot of the legal tech incumbents, they had the benefit of being able to own a lot of this data now, like you don't need troves and troves of this data. It already picks up on a lot of nuance that you wouldn't expect. So I think this shift is actually what's most substantive about this current era of, legal tech and probably why people view this as a renaissance.
00:09:21:21 - 00:09:41:20
Omar Haroun
Absolutely. And I guess building off of that and connecting some of the dots, because you mentioned the profile of certain founders, and you also mentioned how I think like I remember when ChatGPT first came out, there was a prediction that this is just going to really benefit incumbents because they ultimately have so much data. You're a little bit calling that into question.
00:09:42:01 - 00:09:57:23
Omar Haroun
So can you talk a bit about what are the attributes of in your mind, having seen, you know, hundreds of startups now and mature companies? What are the attributes of the companies that you believe have what it takes to win in this, this, this next era?
00:09:58:01 - 00:10:31:11
Dr. Megan Ma
Yeah, I think what's really important is actually it's very relational, this industry. Lawyers are incredible risk hedgers. They love to mitigate risks. And a lot of how they have viewed their own business is through relationships and trust and reputation. And I think that translates in the way that they view vendors as well. I think going back to kind of this evolution, legal tech historically, you know, no one knew about us sort of joke was that it was back office.
00:10:31:12 - 00:10:56:17
Dr. Megan Ma
You know, people just lawyers felt like they needed to use this technology. They were force fed into it. And then they kind of contoured themselves so that they could work with the technology. But for the first time, we're kind of turning that question around and we're saying, what do you need to make your practice better? And, you know, a lot of lawyers are incredibly practice focused.
00:10:56:18 - 00:11:33:07
Dr. Megan Ma
They're building their own individual businesses. And so in this case, a lot of what I see is a successful founder is someone who understands this relational concept. Right. Someone who has a history of understanding that lawyers really care, that you care about their practice. And if you are asking the right questions about what they absolutely hate doing, or what are these kind of little nuances that if you were just able to unlock it, or am I able to kind of tap into my second brain or something like that?
00:11:33:09 - 00:11:54:20
Dr. Megan Ma
That's what really, you know, resonates with them. And, you know, I think, for example, your own history of selling in the legal industry, but also being able to talk with lawyers and understanding almost it's almost like, you know, a therapist, right? Like lawyers feel like they want to be heard. And I think that's kind of what's important about this particular industry.
00:11:54:23 - 00:11:55:20
Dr. Megan Ma
00:11:55:22 - 00:12:29:09
Omar Haroun
Yeah. And I think I mean in my case a lot of it has been pleasant surprise that chief legal officers are really kind of stepping up and in a way that I'd never seen before, like in the last year, are actually, beginning to kind of work with us directly. And I think prior to that, similar to maybe the rest of the world not paying attention to what was going on, there was a little bit of, hey, I should be delegating this to my legal operations person or my discovery person or whomever.
00:12:29:11 - 00:12:40:14
Omar Haroun
It does feel different, and I think there's probably a variety of explanations. I'd love to hear kind of what you think is the evolution of the chief legal officer or why, why they're suddenly getting more involved.
00:12:40:16 - 00:13:12:20
Dr. Megan Ma
Yeah, I think there's an interesting, like, magic that's happened with, people thinking that legal or, just in general even compliance is some sort of, you know, bottleneck. And they're like, oh, God, I have to send this over to legal. And there's suddenly, you know, with technology, they became almost like a positive. All of a sudden it's like, yes, I should send it to legal because I can figure out whether or not this is contributing to the overall strategy of the company.
00:13:13:01 - 00:13:46:17
Dr. Megan Ma
It wasn't just about like, it being a burden. And I think, like, you know, you also see us somewhat pivot in the chief legal officers position in terms of how they see risk. Right. If there is more transparency in how the transactions are being done, then naturally they are not kind of doing a guessing game. We're going off of there like, you know, scars, tissue to tell, and saying like, okay, through running things through this technology, I have some sort of kind of quantitative foundation and bottom line.
00:13:46:17 - 00:14:12:07
Dr. Megan Ma
And now going forward, I can almost gamify how I can make sure that the company is lifted up all together. And so I think, like, it's very interesting to kind of focus on close because in the past, you know, again, like they've found every way to kind of avoid being seen as a cost center. They're the ones that have to make the decisions of when to outsource work to like outside counsel.
00:14:12:09 - 00:14:44:10
Dr. Megan Ma
And now, like, there's just so much more transparency. They know where they can offer their relative value. And then exactly, you know, on the other end, they can almost view in advance, what they should be expecting from outside counsel. And, you know, a lot of what people are talking about in this shift of like, law firm business models and like time and billing is chief legal officers, in a way, are driving that change and saying, like, law firms need to be much more transparent about how much they are charging us per hour.
00:14:44:13 - 00:14:49:18
Dr. Megan Ma
And they have the power now because of these technologies that are available.
00:14:49:20 - 00:15:25:10
Omar Haroun
I'm curious. I remember there's a little bit of some, some studies went viral around the hallucinations related to, to, I'm sure, you know, kind of groups, groups that you're intimately familiar with. We just talk a little bit about this because, I, one of these variances that we had, especially in the beginning of the sudden interest, was it was like people got a little bit too excited and suddenly they were drunk on AI and they actually forgot that there's also, a lot of risk.
00:15:25:10 - 00:15:36:19
Omar Haroun
And I'm just curious how you've seen kind of the evolution of, I mean, our hallucinations are real problem. How real of a problem is there a solution? Kind of. What are your current thoughts on that?
00:15:36:21 - 00:16:02:19
Dr. Megan Ma
Yeah, I think hallucinations. And like every time I see this, I wonder if it's a hot take, but I think it's a red herring, truthfully. It's one of those things where the technology, if you understand the sort of historical background, it's not meant to give you a 100% accuracy. So if you're demanding that the technology has to be 100% accurate, has to be 100% reproducible, you're looking at the wrong thing simply.
00:16:02:21 - 00:16:24:06
Dr. Megan Ma
But on the other hand, if your understanding of kind of how this technology is built, then you will choose your use cases much more wisely. And then hallucination is secondary. And I think in part it's going to eventually, over time, be a cost that you have to account for. And so it's really about moving into an era of what I call division of machine labor.
00:16:24:07 - 00:16:44:00
Dr. Megan Ma
I think a lot of what people view generative AI. And because, you know, there's this mystique and kind of excitement and we it is an incredibly powerful technology that a lot of people make the mistake of doing away with every other technology that exists. Right? It's like, let's get rid of the entire tech stack. But that's not necessarily true.
00:16:44:05 - 00:17:15:00
Dr. Megan Ma
Like I think that, you know, there are some places where older technologies going back to say, the expert systems, if you do want 100% accuracy, that's kind of the technology that you lean towards, because you'll always get the same response over and over again. But if you want kind of technology that is able to tap into almost like your inner conscience or your inner brain and processes and kind of strategy, especially role playing, all of this technology is now possible with generative AI.
00:17:15:02 - 00:17:28:14
Dr. Megan Ma
And so again, like it's really about breaking down. What exactly is the broader problem, how you want to, create or assemble your new team of humans and machines together?
00:17:28:17 - 00:17:45:19
Omar Haroun
I love it, and that's actually a great segue into augmented intelligence. And I would love to just hear, you know, it's it's do you think it's more of an art? More of a science? What what does it really mean to you? And and how do you see that kind of unfolding in the next two years?
00:17:45:23 - 00:18:09:06
Dr. Megan Ma
Yeah, I think like multiple times that we've kind of chatted about this. I see augmented intelligence as a two way street or bidirectional. I think, you know, a lot of people just imagine, like AI as just contributing to the human and elevating them. But I think it's also the other side, too, right? The technology is capable of being increasingly better.
00:18:09:12 - 00:18:32:15
Dr. Megan Ma
The more you kind of work with it. And so it's an era of this, like human computer interaction, human machine engagement, where you're finding ways to, you know, defer certain tasks to, the machines and certain tasks to humans. And it's a call and response. I don't know whether or not we will find that perfect science too, but it definitely is more of an art form.
00:18:32:19 - 00:19:00:07
Dr. Megan Ma
One thing that we've been working on is a going back to kind of the hallucination is really this question of evaluation. And I think for the first time ever, the profession is coming with a reckoning where we are asked, like, you know, what actually do we offer as humans? And especially human lawyers? And we did this exercise where we asked actually partners to define a set of tasks that is an end to end workflow, say in litigation.
00:19:00:09 - 00:19:25:18
Dr. Megan Ma
And we basically gave that tasks out two sets of associates, sets of foundation models and then sets of enterprise tools. And we then got those outputs and we sent them back to the partners in a blind manner. We have x like x number of exhibits. And we say there are three exhibits here. You don't know which is which one is human, one is enterprise and one is a foundation model or ILM.
00:19:25:19 - 00:19:48:22
Dr. Megan Ma
Just rank what you prefer. And I think the results that will come out in our study are going to be quite surprising. I think there is a kind of natural tendency for us to believe that us as lawyers know best, but we're starting to realize that there are certain specific subtasks that the machines have outperformed us in for a long time.
00:19:49:00 - 00:20:09:12
Dr. Megan Ma
And I think what's going to be even more exciting is what's the relative value game between an LLN and actually the enterprise tool. And so all the and comforts that, you know, have argued that we have a special sauce or we have something to show for, I think it will be surprising to them which tasks they really can market.
00:20:09:18 - 00:20:13:10
Dr. Megan Ma
And what actually is the competition out there today.
00:20:13:12 - 00:20:42:19
Omar Haroun
Fascinating. One of their kind of take and I'm really curious to get your perspective on is UX right. We, we all everyone's talking about AI right now. How that gets presented delivered obviously will will make a huge difference. So can you talk a little bit about what are some of the emerging trends maybe, that you're excited about and, and that the audience might, might, might also think about.
00:20:42:22 - 00:21:07:18
Dr. Megan Ma
Yeah. I, I love this question because I think it's really nice that for the first time, the legal community is being asked, like, you know, what is your way of kind of viewing, not technology is just another tool, but almost like a way to kind of, again, do this type of augmented intelligence that you have hand in hand a second brain.
00:21:07:20 - 00:21:41:18
Dr. Megan Ma
And is it from a platform perspective, or is it kind of meeting you where you're at? And I think the latter has been the approach that we've seen historically with legal tech. Everyone's like, we have to make this a Microsoft word plug in. But what we've experienced and from our own kind of prototyping, we noticed that you can't actually unlock the full potential of this technology if you continue to retrofit it with existing, you know, Microsoft tools or any of these existing kind of incumbent platforms.
00:21:42:00 - 00:22:08:17
Dr. Megan Ma
But then it leaves the question, you know, is, is it that lawyers, you know, enjoy asking, you know, legal tech vendors? Can you kind of meet us where we're at and put it into Microsoft Word on Microsoft? 365 because it took us forever to get it to where we're at now that, you know, I had to shape shift everything I do to make my process work.
00:22:08:17 - 00:22:42:18
Dr. Megan Ma
And so I just don't want to deal with that change management or, you know, is it that truly, you know, Microsoft is giving them a workflow that they desperately need? And I think, like until we can prove rightfully that there are other approaches, you're going to have lawyers conflate basically these two that I have to use this because it makes sense for my work versus like, I have to use it because I'm resistant to kind of a new learning curve.
00:22:42:20 - 00:22:58:12
Omar Haroun
Yeah. Yeah. Fascinating. I'm just curious kind of back on this UX point, have you seen much with Boies other sort of mediums? I mean, we talk a lot about text, but, as, as I just kind of curious.
00:22:58:14 - 00:23:22:17
Dr. Megan Ma
Yeah. I haven't seen a ton, but, for us, we've been experimenting. So, for example, that negotiation simulator, we also have a version for depositions. We actually added a voice functionality to it so that you can, you know, if you are training a young lawyer how to negotiate. We made a smartphone capable, or compatible and so on the smartphone.
00:23:22:17 - 00:23:49:12
Dr. Megan Ma
It's actually very annoying to be typing long paragraphs or doing all these different maneuvers. And so we made it such that you can just like argue with the agents on the phone. And that's been something that has been quite nice for, actually partners enjoy it. Oddly, this current generation, they just prefer text. And so they still use the kind of texting to play the negotiation games.
00:23:49:18 - 00:24:13:23
Dr. Megan Ma
What's interesting is that actually the voice functionality pushes them to learn how to capture, like speech, nuance versus textual nuance. And I think that those are a little bit different. You know, I have partners tell me that, yeah, they're associates. They just slack them or like, teams them and they wouldn't even knock on their door even though they're next door.
00:24:14:01 - 00:24:26:00
Dr. Megan Ma
And I think tools like what we're offering and using voice mode actually encourages this current generation to be much more approachable and social when it comes to doing their work.
00:24:26:03 - 00:24:36:14
Omar Haroun
Wow. Fascinating. You've talked a little bit in the past about how legal education is changing and evolving. Can you talk a bit about that?
00:24:36:16 - 00:25:07:10
Dr. Megan Ma
Yeah, I think, well, it's academia is one of those things that if you think the legal industry on the private side is resistant to change, will academia is like deep in the in the perspective of not wanting to change. I think like currently a lot of what we're starting to see is classes that are generally introducing the role of technology in law.
00:25:07:12 - 00:25:42:12
Dr. Megan Ma
But what we're secretly kind of doing here is, say, the negotiation workshop that you see, or kind of the trial advocacy. We've been trying to figure out ways to deploy our technology in that class. So the class naturally has role playing elements to it. And so rather than kind of preparing for the big kind of, negotiation where you do it with your peers, you can kind of have these stepwise improvements that you do on your own, and you can offer it as like homework assignments or sort of instead of readings.
00:25:42:17 - 00:26:08:00
Dr. Megan Ma
I want you to play 4 or 5 games of this particular tool. And I think that's kind of the slow and steady way that we're trying to find. In terms of improving on, their own types of skills, I think with legal education, what people are going to realize is, unlike other jurisdictions, I'm from Canada. So, you know, there's an apprenticeship program that's built in.
00:26:08:00 - 00:26:45:18
Dr. Megan Ma
So like a traineeship, it's a year, you get paid pennies, but at least get to shadow, an experienced lawyer in the US kind of world, like, it's just a baptism by fire. You graduate from law school, you write the bar, then you're thrown into a law firm. And I think without this kind of training year where you're encouraged to fail and fail fast, but then learn from it, you actually kind of instill a culture of fear in a lot of the junior associate, and especially if you're continuously taking away, like, the tools that you've historically said were going to be useful.
00:26:45:18 - 00:27:13:23
Dr. Megan Ma
So clients, for example, requesting that they don't want associates on the case because they don't want to, you know, have your associate like, be trained on their own client matter. That's kind of like creating a problem with young associates, kind of picking up on that legal nuance. In you know, the litigation space, you know, if you're in a contingency fee model, we're starting to see partners saying like, well, we don't get paid if we don't win.
00:27:13:23 - 00:27:38:03
Dr. Megan Ma
So I'm not going to have my associate on this case. They can just observe from afar. So observing is not really learning in any way. And so we're really trying to figure out if you're bridging this gap that's going to be exponentially larger. How do you actually kind of culture in into the law school itself. Technology using generative AI to find ways to, pick up on these practical skills.
00:27:38:05 - 00:28:01:12
Dr. Megan Ma
The other thing that I started to notice is notebook LLM was really popular. It had a viral moment of being able to generate these podcasts. It was kind of fascinating because in the beginning there's so much resistance from professors. And then later on hearing fictional agents tell them about how amazing their work is apparent can sell very well.
00:28:01:14 - 00:28:12:13
Dr. Megan Ma
Turns out that like papers from like, 30 years ago suddenly have a renewed excitement. So I think that, for example, at Stanford, we started to think about using notebook LLM for our students.
00:28:12:15 - 00:28:34:20
Omar Haroun
That's amazing. Never underestimate the power of flattery as a person. Everybody, who's who's paying attention to legal technology knows Codex and now increasingly knows you. Megan. I think one thing that'd be great is to hear a little bit more about your own story, how you got to where you are. Any any things about you that may not be, you know, obvious as someone just reads a bio about you?
00:28:34:22 - 00:29:01:23
Dr. Megan Ma
Yeah, sure. So, I actually had one of the windiest road spot. I think others, like, probably when I imagine, I didn't really have that much of an interest in technology itself. I actually started with an interest in the medical law space. But at some point, kind of I stumbled upon, you know, reading about the launch of a bunch of this medical AI back in 2017.
00:29:02:00 - 00:29:29:14
Dr. Megan Ma
And it was, you know, this company called Babylon Health that kept touting that they were, you know, at 92, 93% accuracy in terms of being able to diagnose. They were like, I don't I watch a lot of house. So they were like, oh, you know, we're like house level or something like that. And then they kind of compared themselves against, you know, Oxford trained and Harvard trained physicians with 30 years of experience saying they're only at 85% accuracy were at 92.
00:29:29:16 - 00:29:55:21
Dr. Megan Ma
And I was kind of stunned by this comparison. And so I started digging deeper because the question that I had, you know, obviously being trained in law was, well, what happens at the 8% misdiagnosis with doctors? There's a transparent accountability scheme here with these tools. There's no one to turn to. Do you turn to the engineer? They can't really tell you exactly why there was this misdiagnosis.
00:29:55:23 - 00:30:18:23
Dr. Megan Ma
And so I started getting fascinating, fascinated by, you know, how does law kind of pivot around these questions of technology that come up? And then later on, I wanted I was, you know, working in the government, and I worked in the Ministry of Transportation in Canada, and I was assigned to work in an inter-governmental group on self-driving cars.
00:30:19:01 - 00:30:44:06
Dr. Megan Ma
And funnily enough, you know, we were on a lot of calls with California, and they kept telling us how much of a success story they are. And they were like, you know, if I were to give you two pieces of advice, kiddo, like, as you know, America likes to see Canada. They said the first thing is you have to ensure that there are little selectors between lanes.
00:30:44:09 - 00:31:13:04
Dr. Megan Ma
So that these cars can see, you know, between lanes. Much better. The second thing is, is that you need to change all of your traffic lights such that they are now led. And the reason behind the LED lights is because, again, you know, they're much better at helping self-driving cars with signals. And we kind of took back this advice and we thought about it and we're like, wait, this absolutely wouldn't work in Canada.
00:31:13:06 - 00:31:40:03
Dr. Megan Ma
And really boiled down to something as simple as weather. And in Canada, it snows. And so when it snows, it will absolutely cover, you know, the little reflectors between lanes. So that would be rendered useless. I think the worst part is, is that if you, you know, invested in a multibillion dollar project to change all of the traffic lights, you're going to run into a problem where all of these lights are going to collapse at the first snowstorm.
00:31:40:03 - 00:32:08:17
Dr. Megan Ma
And the reason is because the current traffic lights, they emit a little bit of heat, meaning that when it snows, it'll melt the snow off the traffic lights. LED lights don't melt the snow. They don't emit the heat. And so even if you're kind of trying to, you know, translate a new technology and the lessons that you learned in your own jurisdiction, they simply can't work from very few, you know, seemingly innocuous factors like weather.
00:32:08:19 - 00:32:42:09
Dr. Megan Ma
And I think that these two kind of lessons in medical, I found in self-driving cars, it really informed how I see technology in the broader legal field. I think when it comes to, say, the self-driving cars, what it taught me is that translation across jurisdictions and across even the legal field, across different types of law, it is a very, very granular process, and it's about, again, understanding something down to the individual lawyers and down to their individual practices when it comes to like medical AI and questions of accuracy.
00:32:42:12 - 00:33:04:18
Dr. Megan Ma
I realized that like, you know, how much of the practice is really about accuracy, like at it, when it comes down to it, you're not going to sell a lawyer because of how accurate they are in giving you their advice. It's about how they formulated that argument. And in the same way, with medical care, it's about how they shared their expertise to you when you came to them with an illness.
00:33:04:19 - 00:33:24:05
Dr. Megan Ma
And so I think, like for me, I look back as if like, you know, these lessons all made sense. But at the time, I was just feeling around and trying to see, like, you know what interested me and I, I think, like, at the end of the day, it came down to technology was fascinating to me and then how it intersects with law.
00:33:24:09 - 00:33:46:01
Dr. Megan Ma
And so I just state, I think this is something I asked before, but, you know, this is kind of the second time around for you. And I thought previously, like, is there any deja vu? But to me, I think I'm most fascinated by what excites you again about the legal tech industry. You know why legal?
00:33:46:03 - 00:34:16:07
Omar Haroun
Yeah, definitely. So for me, it's it's it's really two things. One, from a mission perspective, I went through this, period of kind of reflection and not entirely positive, I would say, where right after selling my last company, you know, we had a good outcome and our investors and team had a good outcome financially. But but really, the we actually changed the industry in any way.
00:34:16:07 - 00:34:38:09
Omar Haroun
No. And that was kind of disappointing. And I think also just realizing that we never even really had much of a mission. I mean, our mission was to build a business. But when I got closer to the legal industry and spent seven years kind of in the trenches getting to see some of the realities up close, they weren't all, nice, nice things to kind of witness.
00:34:38:11 - 00:35:03:05
Omar Haroun
And I really started to see the connection between the system, which kind of at its core is propped up by a lot of perverse economic incentives and almost kind of a, guild that's keeping, the law more complex than it needs to be. And, you know, there is sort of a consequence of that, which is that I think there's really two groups that are harmed.
00:35:03:05 - 00:35:23:21
Omar Haroun
One is businesses who, to your point earlier about seeing legal as a cost center are actually constrained and have to pay this tax in order to innovate. And if you're big enough to afford it, that's fine. But, so annoying. And if you're a small company, I mean, even doing like a series A financing is like $50,000, right?
00:35:24:03 - 00:35:44:15
Omar Haroun
And so it's just crazy, like for, for a small startup with, with almost no revenue and history to be able to afford these legal expenses. So I think I think on the business side, that's the situation on the much more, I would say sympathetic side. You know, you have people, consumers where a single legal event can still bankrupt the average American.
00:35:44:17 - 00:36:06:19
Omar Haroun
And we still have a system that, most people who need lawyers can't afford them. I mean, the research on this is staggering. It's like 90% of people who actually have a legal claim can't afford a lawyer. And and yet we have AI now, you know, as Armand said in a previous episode, you know, it's it's really capable of bringing abundance.
00:36:06:21 - 00:36:24:21
Omar Haroun
If it's really hard, it's in the right way. And from a pure technical perspective, well, there's a lot of nuances you have to get right. At the end of day, this is a text based. I mean, we are analyzing language, generating language kind of at the core. So so there's this like obvious opportunity to bring the technology to this space.
00:36:24:21 - 00:36:47:12
Omar Haroun
And the reasons that it hasn't been adopted aren't really, entirely I mean, they're logical, but but but they're, they're, they're kind of the wrong reasons. Yeah. Right. Like it's it's, it's more a matter of, I think a it's very complex. But to the extent that you start to unpack it, basically there's people are making a lot of money off of a system that's broken.
00:36:47:13 - 00:37:17:02
Omar Haroun
So I, I think for me, I, I've, I've, I've really primarily been motivated by that this time around. And it's very empowering. And I think thankfully we've had patient long term investors who have very kind of, you know, large amounts of capital to deploy towards, like really fixing the system. And I think they're also genuinely inspired by that kind of profit and purpose alignment, not not just building a big business, but also helping to kind of reshape whole industry.
00:37:17:04 - 00:37:42:16
Omar Haroun
I think the the second factor for me was, timing. And what I mean by that is I've, I've really not only thought about my own business, but also been involved in, like, dozens of other startups, one way or another, just helping people, investing, advising. And I actually think, timing, market timing is the most important factor in your success and the one that's the most out of your control.
00:37:42:18 - 00:38:03:13
Omar Haroun
So you can have the best technology in the world, the best team in the world. If you're a little bit too early or a little bit too late, you're going to have a really hard time. And so when I was, you know, kind of reflecting and just thinking about what do I do next with my life and, some of these questions, I just started interviewing a lot of chief legal officers who had gotten to know a little bit previously and did relationships with.
00:38:03:15 - 00:38:26:23
Omar Haroun
And I was just shocked by how suddenly, you know, kind of to the stuff we were discussing earlier. It's like generative AI transformation is a top three priority for the CLO, in a way that I never had been. At the same time, the risks and the lack of comfort with tech companies and even the incumbents who are mostly disappointed, historically, or they hadn't really been involved with directly.
00:38:27:01 - 00:38:40:00
Omar Haroun
I mean, it just like there's kind of this, this magical moment of if we actually did bring a private, secure, trusted version of the technology and we were willing to kind of thing from first principles, there's there's a really good opportunity there.
00:38:40:01 - 00:39:07:01
Dr. Megan Ma
I really like kind of what stood out to me in that response is the empowerment side of it. At Coda is actually our mission is empowerment through legal technology or technology itself. And there's a kind of great synergy there. And I think like, it's I really see kind of why chief legal officers, are a great user, kind of or a user base for this.
00:39:07:01 - 00:39:37:20
Dr. Megan Ma
I think, like a lot of lawyers that I talk to, many of them are like, you know, if you think about basic supply and demand, like there will always be demand for us. I it's just funny because they can say that because we haven't had real metrics around what is quality legal work. And once we have the tools that we give to their own clients to identify what is quality, then they can't, you know, that supply demand is going to shift dramatically.
00:39:37:22 - 00:39:57:02
Dr. Megan Ma
I always think that, you know, to your point, I think a lot of the technology is already getting there. But adoption is always a structural hurdle. It's never to do with like the technology itself. And I think a lot of legal professionals will say, oh, we have to like reach a certain level of trust with the technology.
00:39:57:02 - 00:40:18:09
Dr. Megan Ma
I think that's sort of a fiction. Yeah. Naturally, like, I think a lot of people say no UI is the best UI kind of thing. I think over time, people will fade away from seeing generative AI as like technology or some sort of like tool that they have to confront. Instead, it's going to be an organic, kind of like hand to hand relationship.
00:40:18:11 - 00:40:39:14
Omar Haroun
Fascinating. Well maybe it maybe that's, that's, that's a good segue. Let's, let's kind of close our eyes and imagine the world, you know, ten years from now. And if based on what you're seeing now and just the pace of innovation, what do you see the world looking like, the legal industry in particular, looking like.
00:40:39:16 - 00:41:08:01
Dr. Megan Ma
Yeah, I think this is probably our moment today. Of legal experiencing for the first time, the empirical turn. I think other professions have seen it in the past. And so if we're thinking about like five, ten years from now, I imagine that lawyers won't talk about technology as if it's a phenomenon. And so, you know, you don't go to accountants and say like, oh, you know, like how do you crunch your numbers or like, give me the exact tools.
00:41:08:01 - 00:41:34:11
Dr. Megan Ma
Right. But lawyers now kind of detail all of the software that they use and all the technology they use as if it's something to keep at arm's length. And so kind of to the point that we've been touching upon this question of augmented intelligence, I think they're just going to not even discuss, machines because everyone will have reached a point of being acclimated with a certain tech stack, possibly Judea.
00:41:34:13 - 00:42:01:10
Dr. Megan Ma
And so I think, like you know, to me, the best way is to almost view the technology as if it's not there because it's already absorbed into your practice, because everyone will see that you're going to have a competitive edge. And if you don't use it, then you'll be seen as like basically the grandpa in the corner that like, you know, is still shouting at people to get off the lawn.
00:42:01:12 - 00:42:16:09
Omar Haroun
I love it. A great, a great image for us to, to end with. Well, thank you so much for for joining us today and really excited about, continuing to kind of watch your, your own growth and Codex is, major impact on this, on this industry and space.
00:42:16:11 - 00:42:17:02
Dr. Megan Ma
Thanks for having.
00:42:17:02 - 00:42:17:22
Omar Haroun
Me. All right.