a16z Podcast - The Robot Lawyer Resistance

Episode Date: March 1, 2023

What happens when you don’t have the resources to fight for your rights? Whether it’s a major medical bill or a more minor parking ticket… even if you were in the right, is it easier to just… ...pay?Well, Joshua Browder's answer is Do Not Pay! He’s built a thriving company of the same name, that helps consumers “fight corporations, beat bureaucracy, and sue anyone at the press of a button.” Through the power of technology, DoNotPay has resolved over 2m cases successfully and in this interview we get to hear where Browder hopes to take the organization, including his first-hand account of his recent plan to bring the robot lawyer into the physical courtroom. Topics Covered:00:00 - Introduction01:51 - DoNotPay’s origin03:09 - Surprising laws05:49 - General counsel for consumers06:54 - The role of AI and technology09:26 - Ensuring quality counsel13:19 - Prioritizing features14:43 - Corporate pushback17:11 - Reshaping the legal system20:28 -  Entering the physical courtroom 24:50 - Cross border possibilities26:14 - Looping humans into modeling29:01 - Deflationary impacts30:05 - Recent pushback33:13 - Is AI already writing law?35:15 - Why is law so complex?37:54 - Entering the Supreme Court38:38 - Lawyer support39:47 - Looking forward43:07 - Building horizontal vs vertical45:28 - Does this law exist? Resources: DoNotPay: https://donotpay.com/Find Josh on Twitter: https://twitter.com/jbrowder1 Stay Updated: Find a16z on Twitter: https://twitter.com/a16zFind a16z on LinkedIn: https://www.linkedin.com/company/a16zSubscribe on your favorite podcast app: https://a16z.simplecast.com/Follow our host: https://twitter.com/stephsmithioPlease note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. For more details please see a16z.com/disclosures.

Transcript
Discussion (0)
Starting point is 00:00:00 I imagine a world where you didn't even know that something happened, robot lawyers were fighting to you in the background, and almost immediately they get you the re-findled something back. Today we're talking about the fascinating intersection of law and technology. So the law is a set of rules that a group of people agree upon, to regulate a specific region, often through the application of penalties if people break such laws. So, for example, in many regions, it's against the rules to steal something from another person. and if you're caught breaking such law, well, a penalty is applied.
Starting point is 00:00:33 But what happens when you don't have the resources to fight for your rights? Whether it's a major medical bill or a minor parking ticket, even if you were in the right, is it easier to just pay? Well, Joshua Browder's answer is, do not pay. He's built a thriving company off the same name that helps consumers fight corporations, beat bureaucracy, and sue anyone at the press of a button. Through the power of technology, Do Not Pay has resolved over 2 million cases successfully.
Starting point is 00:00:57 And in this interview, we get to hear from Browder himself about where he hopes to take the company, including his first-hand account of his recent plan to bring the robot lawyer into the physical courtroom. You're not going to want to miss this. As a reminder, the content here is for informational purposes only. Should not be taken as legal business tax or investment advice or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund. For more details, please see A6C.com slash disclosures. I am so excited for this conversation today. I've been seeing you post on Twitter. I've been seeing news articles everywhere about what do not pay has been experimenting with.
Starting point is 00:01:50 And so we'll get into some of those timely topics. But first, let's start out with where do not pay originated. So how to do not pay start and what really inspired this company? So it's no secret that I'm a terrible driver. And when I started driving at age 18, I got about 30 parking tickets between the ages 18 and 20 in London and also when I moved here to study in college. And I couldn't afford to pay these really expensive tickets. And I learned something remarkable, which is if you know the right things to say,
Starting point is 00:02:20 you can save a lot of money. And I got this reputation among my family and friends is the person who can help with parking tickets. And it wasn't long before people were just messaging me just pictures of their tickets. They wouldn't even say anything. I didn't want to copy and paste all of these documents. And so I thought, why not build an automated solution
Starting point is 00:02:39 with templates where a user can select an option and then a defense is generated and sends the right place? And I really just did it for a few family and friends. and I never could have imagined that it would go globally viral because people hate parking tickets. And that's what made me realize that the idea of helping people with technology and consumer rights is bigger than just tickets.
Starting point is 00:03:01 And so I spent the past almost seven years now expanding Do Not Pay from parking tickets so many other areas of the law. Well, yes, many other areas of the law. Your website has, I think, over 100 different use cases or applications that you help your customers with. and I just want to call out a few and get you to quickly explain what this is, how it works,
Starting point is 00:03:24 because I think many people are maybe familiar with Do NotPays, parking ticket, support, things like crafting a cease and desist letter. But let me just give you a few examples that I was surprised by. So one of them that I saw on your website is free raffle tickets. What does that mean? So there's this obscure law in America that says that you can enter any competition for free. So you hear on the radio, you can pay to enter, but they always bury in the terms of service that if you mail an obscure letter to an obscure address, you can enter for free.
Starting point is 00:03:56 So we've built a product that automates free entry into any competition. So you just paste in the terms of service, or you can even just choose from competitions that are bought scrapes on websites, and it enters it for you for free. And this is the thing with Do Not Pay. There are so many laws that help consumers that no one has the time to actually follow through with. And that's a great job for technology. And it empowers those people through technology, right? I feel like there are people who know about these loopholes within raffles, for example, but they're just manually entering into these things without the knowledge or without the support
Starting point is 00:04:30 of technology. Let's go through one or two more just to kind of give people an understanding of the vast range of areas that law can be applied. One of them was egg donor rights. What do you mean by that. There are all these rights around egg donors and making sure you get compensation for expenses and things like that, which is important. And also on the other end, there's compensation opportunities for donating that sort of stuff. And we've automated kind of signing up for that if people need extra money. It's a very controversial product, but we're happy to give people more money. I'm sure there are many controversial products that you facilitate. What about this one? This one's not so controversial, but I thought it was interesting. Find my lost pet. Every city or most
Starting point is 00:05:17 cities have these forms where if your pet is lost, you can fill it out. And if the pet is found on the street and it walks towards a fire station or a police officer, they'll deal with it. And then they will put it into the system to see if someone's reported it as lost. And this is another area where people don't have time to fill in these obscure government forms. The same is true for airport lost property forms. We've automated all of these lost property forms. We've automated all of these lost property forms at major airports. So Do Not Pay is really about saving people time with obscure consumer rights laws
Starting point is 00:05:47 and also government services. Yeah, I like that positioning. And another positioning that I've heard you say is basically the general counsel for consumers. And I love this idea where the average everyday person doesn't necessarily have the funds or the time to represent themselves effectively and you've been able to use technology to empower these folks.
Starting point is 00:06:06 So let's fast forward. You started seven years ago. where is do not pay now? Like how many customers are you servicing? How many people are actually using these tools, as you said, as a general counsel? So we've submitted over two million cases and we have hundreds of thousands of active subscribers at the moment. What's really exciting is that a lot of this stuff is not necessarily the obscure products you mentioned, but getting serious refunds from companies. So refunds are our most popular product, canceling subscriptions, Some gym memberships require you to send a signed letter
Starting point is 00:06:39 and mail it to an address just to counsel. Suing robocallers, getting discounts, these kind of high-ticket items that people don't have time to wait on hold for hours just to get a couple hundred dollar refund. And so that's a good job for software as well. So tell me a little bit more about that particular use case where you are helping someone get a refund or get access to something where it typically would take hours
Starting point is 00:07:04 of someone sitting on the phone. How are you actually facilitating that with technology? So our motto is the squeaky wheel gets the grease. So if you are on a flight and the in-flight Wi-Fi doesn't work and you've paid, it's now $50 for transcontinental in-flight Wi-Fi and it doesn't work. Our software will generate a letter and send it off to United Airlines. And it will cite FTC statutes and be really aggressive. United will receive that letter and think it's not worth it.
Starting point is 00:07:32 And so 80% plus of the time, they just grant the REITC. fund. And that's what's amazing is just with a template letter, we've managed to achieve all of this success over the past few years. But what's really exciting is in the past kind of six months, we're actually using more and more true AI versus decision tree templates. Tell me more about that, this differential between maybe what people did in the past, which was a human lawyer drafting up a letter to, I'd say a second step, which is using technology, not necessarily AI, but really to take different pieces that are necessary in crafting them, almost like a template. And then a third generation, which is using AI, which is not just a template, but really
Starting point is 00:08:15 bringing in all of the prior examples for a particular letter and crafting something net new. How do you think about which use cases require step one, step two, or step three, or how are you assessing where to apply some of these new technologies? So we started at step two. And so what we did was we said, all of these letters that you had to pay a lawyer $300 for, we're going to automate with document automation, which is a very simple technology, but it's something that hadn't been done
Starting point is 00:08:44 for these types of letters we were doing. What's really exciting about step three is that sometimes the companies respond and the government's respond, and step three with AI allows us to respond back instantly. So instead of helping someone with a $100 dispute with Comcast, we can now work on $10,000 medical bills. because Comcast or United might let that $100 go,
Starting point is 00:09:08 but no hospital is just going to let $10,000 go just because of one letter. And so you really have to have that conversation back and forth. And so what we're working on now is actually communicating live on medical bills specifically, and I can go into the laws we're using around that. That will be kind of our first major AI product in Do Not Pay. Please tell me more about that
Starting point is 00:09:29 because something that I would love for you to address is the fact that some people may push back on this idea of having a robot lawyer because some people might say that actually having a mediocre lawyer or a bad lawyer is worse than having no lawyer or representation because you might get yourself into more trouble. I don't know if that's a case with the medical bill example, but tell me more about how you're using these AIs to really have this back and forth and ensure that it is the high quality back and forth that you would want
Starting point is 00:09:59 if you were actually enlisting counsel. Do not pay as automated consumer rights, and unfortunately, lawyers don't typically get out of beds for consumer rights. I said that on Twitter that they don't get out of bed at all, and everyone's crazy, because they're saying, I want sued a client for $500, I get out of bed for consumer rights, and so most lawyers focus on serious issues. And because of that, these corporations' governments know that people can't fight back. And these issues aren't rocket science. If the in-flight Wi-Fi didn't work, you're entitled to a refund, and there's very little that technology can do to change your case. And so it's really about just jumping through the
Starting point is 00:10:36 hoops, saying the form letter and sending off to the right place. With medical bills, we haven't yet launched it, and we're testing it, and we're trying to be very thorough to make sure it's accurate before we launch it. But the way it works is there's something called the No Surprises Act, which is a brand new law passed in 2022. And it requires hospitals to do two things. The first is, to publish their prices. And a lot of these hospitals, when they were complying with the law, they published their prices in like obscure PDFs and Excel documents on their website. And so what we've done is we've had GPT scrape all of these websites,
Starting point is 00:11:12 create master Excel of 1.4 million rows because they don't name these treatments properly. They just put obscure codes and standardize it among all hospitals in the U.S. So that's the first step. And the second step is this law requires hospital. hospitals to provide a good faith negotiation to provide discounts on medical bills. And so the way it works is say you upload a bill, it takes what's in the bill itemized, compares it to this master list of nearby hospitals offering the same treatment. And if any item is more expensive than a similar hospital, it crafts it in this latter to say that I'm asking for a good faith discount
Starting point is 00:11:52 because a similar hospital is offering the treatment at a cheap rate. And it's not even a different a hospital. Sometimes the same hospital is offering a treatment, but for insurance, it's like 50% of the cost if you're paying the cap price. And so all of this stuff is not rocket science. You can have a lot of impact with just a letter. And a lot of lawyers say, well, if you're being accused of a serious crime, maybe AI is not a good idea. And I think that's true. But sad to me for fighting back against corporations, we haven't seen much go wrong. And I think it can be a net positive it's the people. Yeah, I want to get into this further conversation around what AI can and can't do and also your venture into the physical courtroom. But I think what you called out there is so
Starting point is 00:12:33 important, which is just in many cases, we would hope that the law is applied uniformly and that it's applied in a seamless way where you don't have disparities and information. But the example you gave is perfect in illustrating that sometimes it is just having access to counsel or sometimes it is just having access a certain piece of information to know what laws exist or what laws don't exist or in the case of in-flight Wi-Fi. You're right. If something doesn't work, you should very easily be able to get your money back. And so with that said, I want to understand from you more about how you're thinking about evolving the different services that you offer. I mean, you already offer, as I mentioned, 100 things on your website. So how did you decide, for example,
Starting point is 00:13:18 next up is medical bills. We're addressing that. Or how do you decide, oh, actually, we want to go into the physical courtroom. Like, what are you doing to understand where your priorities should be? We're just trying to create value for people. So we look very carefully about the average amount that do not pay customer saves on a yearly basis. And we're trying to 10x that.
Starting point is 00:13:37 So instead of it being a few hundred dollars, a single medical bill could be $10,000. And that's where I think AI comes in, really raising the value of the cases that the robot lawyer can pursue. We've tried things that haven't worked well because it's too emotional. AI is not too good at dealing with emotions.
Starting point is 00:13:55 So many years ago, actually, before I turned Do Not Pay into a business, I tried to use Do Not Pay for asylum, like to automate the asylum process. But with these cases, sometimes people are crying in the courtroom, and that's something that software can never really touch because you need the human touch in those sorts of cases. That's fascinating. When you're talking about the asylum case, for example, you're saying people need to go into the courtroom and vouch for themselves.
Starting point is 00:14:19 Is that what you're referring to? Yeah, and it's all about emotion. It's about the judges. And of course, there's a body of law that they rely on. But it's beyond the paperwork. With a bill or with United Airlines, it's really about the paperwork. But once you start getting into emotions, like real divorce cases is another example. They have to have like marshals in the courtroom because people might even fight in the courtroom if they weren't there. And those sorts of cases can't really be touched by software. The corporations that you're fighting against, how have you seen them respond to the work that you're doing? Has there been a pushback from these companies saying, no, no, no, no, like you're kind of messing with the system that we've built. You're giving people information or access to things that we kind of banked on them not having access to. It's an arms race. So we saw this from the very beginning at Do Not Pay, where we had bots do clicking on big company websites, and they would have anti-bots to try and identify
Starting point is 00:15:14 buy our bots and shut them down. And the way we got around that is we have the Do Not Pay app and actually uses the IP of the user. So it's almost like lots of distributed nodes fighting these companies and they haven't been able to stop that. On the AI front, the big companies are using AI now to stop consumers even outside of Do Not Pay. And that's one of the things I think is sad about new technology. It typically gets in the hands of massive corporations and governments first. So our role is give power to the people first. So one example is we have a bot that we're using internally to negotiate Comcast bills. Comcast have an online chat and it generates these instant responses to chat with Comcast. And it's obvious that Comcast is already using AI. And so our AI is chatting with
Starting point is 00:16:00 Comcast AI. And on the negative side, it's kind of sad that consumers had to chat with AI to get refunds. On the positive side, at least now the AI is doing the work on both sides and saving the humans from the bureaucracy. And are you finding that you are able to keep your AI up with the AIs that are written by Comcast, for example, where they are actually winning those battles? Yes, and I think it's two reasons. The first is that I feel like I have the best job in the world, and our team is way more motivated than the average big company engineer.
Starting point is 00:16:33 And then the second reason is the kind of drop in the bucket. Do not pay is helped a lot of people, but if you kind of divide our use cases by our customers and the number of companies out there, it's just a small tax that they're having to pay. Maybe if we one day get to 50% of their revenue, they'll actually care properly. But at that point, it will be so kind of huge that we'll have other issues to worry about.
Starting point is 00:16:55 I mean, that'll be a good problem to have. But I think it's actually a great question, this idea of if you are able to expand and really give consumers the legal counsel that they need, or in this case, like to be able to fight back against fines or charges that they should, be paying if that can expand or scale to a substantial part of the market won't these companies have to rethink the way their businesses work won't new laws have to be written i mean one simple
Starting point is 00:17:23 example of this that i think of is parking tickets where up until the pandemic you could kind of count on the fact that if you went and fought your parking ticket most people wouldn't bother with this but if you did the police officer might not even show up but then when covid hit all of these parking ticket cases were online. And that was a totally new dynamic. More police officers are showing up. More people are showing up to fight their cases. And so the change in the system required people to rethink, like, are we doing this properly? And so have you seen that at all? Are you expecting that where companies are going to have to rethink the way that they're engaging with customers? Yeah, I think this is one of the biggest risks to do not pay. What if all
Starting point is 00:18:03 the problems in America get solved? Fortunately, America, I mean, it sounds hypocritical with my accent. but America has a lot of work to be done and I think it will take a long time before everything is perfect. And if it is, we'll be living such a good life. On the courtroom front, it's really interesting because I think COVID has accelerated the legal industry by 20 years.
Starting point is 00:18:25 A lot of the things stopping a company like Do Not Pay were these bottlenecks that were introduced sometimes over 100 years ago. You have to show up to fill in this form. You have to have a wet signature for this document. So one product we have is we all, automate the filing of 83B elections, which is an obscure tax form for startup employees. And up until COVID, the IRS required a wet signature on that document. But during COVID,
Starting point is 00:18:51 they relaxed and said you can have a digital signature, allowing us to build the product. And the same is true with these Zoom court hearings and sometimes filing of documents online. And that's actually moved us forward. And what's exciting is we continue to be forward in 2023 with all these relaxed regulations, but traffic tickets now are going back into court. So we're having all the benefits with few of the disadvantages. And is there pushback at all from the people who are facilitating these courtroom traffic ticket cases, for example, where they're like, I don't really want to bring this back to the physical world? Like, why can't we stipulate this digitally?
Starting point is 00:19:27 These judges are very old, very anti-technology. And even in the cases that do allow Zoom, I'm a legal nerd, so I watch a lot of, like, legal cases and view transcripts, and they say things like, nice of you to show up today when the Italian is on Zoom and they say something like, oh, yeah, I'm sorry, I'll state for another case. And so they're already being prejudiced against Zoom, which is not a positive outlook. But I think it just shows how the judiciary and the legal industry is being controlled by very old incumbents that might not have the most forward view on technology. To a degree, I'm not surprised, but in a way I am surprised. because the amazing thing about technology is it often automates or gets rid of the boring, mundane aspects of a job. And then you can imagine how if these judges didn't have to spend all day in court talking about parking ticket cases, they could apply their wealth of knowledge to other maybe more intriguing, interesting, diverse cases.
Starting point is 00:20:26 But with that said, I think this is actually the perfect segue. I was going to wait a little bit, but let's move on to what you were hoping to experiment. with, which is actually bringing your robot lawyer into a physical courtroom. So let's start there. What was this project? Yeah. So in December of last year, I made an offer on Twitter and I said that even if you lose, do not pay, we'll pay the ticket, but we want to use AI to represent you where for a speeding ticket, a user would go into the courtroom, have these bone induction glasses, which are basically like air pod version of glasses. And the AI would whisper in there are telling them exactly what to say to get them out of their ticket. And we had about 300 people
Starting point is 00:21:09 stepped forward and offer their cases. And we accepted two of them. But that's when things went crazy with the pushback from the legal industry. So what did that pushback look like? Lawyers were writing letters to state bar associations. One lawyer called up every major state bar association in all the major cities. And what specifically about that experiment was against the rules in their eyes? The legal industry is very protectionist. So to even become a lawyer, you have to pass the bar exam, take a test, have a government stamp of approval, even for something as minor as a traffic ticket. And their view was that even though the laws weren't written,
Starting point is 00:21:47 they could never have imagined that when the laws were written in the 1800s and 1900s, that AI would be practicing law, this applied to that situation. But our view it do not pay was that it didn't apply to that situation because in these laws it says a person can't practice law without passing the bar and AI is not a person, it's a technology, and the person is just representing themselves with the help of AI. But they obviously don't seem to buy into this.
Starting point is 00:22:13 I also want to understand your perspective around, you said this one person seemed very adamant, very, very emotional about this idea of bringing AI into the courtroom. What do you think that is a reflection of? Why was there such effort to stop this experiment? Lawyers love rules, and they love accusing people of breaking
Starting point is 00:22:34 the rules. And I think they're scared of this technology. It's like the dinosaurs suing to stop the ice age. Lawyers really don't want this technology to move forward that much because they realize that a lot of their job is copying and pasting documents. Not all lawyers are like that, but some of them are. And if AI can just do that and replace them, it puts a lot of them out of work. And so it's a combination of loving rules, being worried about their profession, and also just disliking these young kind of CS students trying to take them down. I'm not part of the legal establishment and I'm trying to fight against them. So I also think there's like a cultural mismatch.
Starting point is 00:23:14 So use the term ice age there, which makes me think that you think this is somewhat inevitable. Like even though there is pushback, even though you were threatened with six months of jail time, that this may be a blip in the journey, which eventually will result in technology being applied to the legal practice. maybe in a substantial way. Is that an accurate depiction, or how do you see this progressing past this one instance? Yeah, I mean, so a Colombian judge recently wrote one of their opinions using chat GPT. Lawyers themselves are using this technology, but instead of passing the savings onto the client, they're keeping the rates the same and just pocketing the savings for themselves. And so our viewer do not pay is that customers deserve the savings. They
Starting point is 00:23:56 deserve to access their rights very cheaply, and you can just cut out. the middleman with these lawyers taking huge sums just to write a letter. I think a fair pushback on that would be, well, that's true. It'd be great if the savings were passed on to the consumer, but does the consumer who has not passed the bar as an example have the discretion to understand if what the AI is returning is actually something that's going to represent them well? So is that maybe even a fair question, or how would you respond to that?
Starting point is 00:24:25 Everyone points to when an AI makes a mistake, but we forget that humans also make mistakes. Their bar associations actually exist to police lawyers. There are lots of lawyers who steal money from their clients. Most lawyers are expensive, but they also steal money on top of that. Somewhat lawyers unfortunately have drug and alcohol problems. And AI isn't perfect, but at least it's objective and it can't get drunk the night before the case. That's a good perspective. I think another amazing thing is that the technology can be cross-border as well. So typically, a lawyer is specialize in one very particular area, maybe in one jurisdiction. You tell me a little bit more about how you're thinking about the potential for this technology to support people, not just in your
Starting point is 00:25:09 region, but really anywhere. Yeah, the benefit of AI is it can read 10,000 documents and produce an answer in seconds, and not even the best human lawyer can do that. So I think the example you're talking about is a consumer rights issue relating to a timeshare. We had a elderly consumer in Boston, they got suckered into a timeshare actually in Mexico and do not pay got them out of it because there's a five-day law in Mexico to cancel timeshare contracts. And so things like that where you can help consumers in Boston for issues in Mexico, most lawyers in Boston don't know about Mexican law, but AI does because you can just feed it all the timeshare cases and it comes out with a way out. And the same is true actually
Starting point is 00:25:51 for Comcast, these generic AI models like DaVinci from OpenAI, They're not actually that good at the law. What we've had to do at Do Not Pay is you feed it documents. So we say we fed at all of these outcomes from the templates in many years of operating. And you say based on these thousand documents, write X. And then it becomes much better than just a generic usage of any of the models. Something that OpenAI has also done, though, is had humans as part of the training process to really hone it in. Are you also doing that?
Starting point is 00:26:23 Are you bringing in lawyers who are specialized in the, these areas to take a look at the output from your models to hone them in or improve them? Yeah, we have some of the best lawyers in the country helping us like at Wilson-Sensini and other firms. The AI has been very problematic for us. So there's really two problems with it. The first is that it's dishonest. So with Comcast, for example, it says things like, I've had three internet outages in the past 24 hours. And that might be a good way of getting a refund from Comcast, but it's not true and it can court liability issues. And And in court, if you lie, that's a crime and send people to jail.
Starting point is 00:27:01 The second issue is that it talks too much. With Comcast or even in a courtroom, there are some questions and human speech that don't require a response. It's like a rhetorical question. So imagine the judge says something like, hold on, let me take a look at the case. The AI models would often say, okay, thank you. And by the way, I'm innocent for like these six reasons. The judge will just get annoyed. You have to build other AIs to decide whether to say something.
Starting point is 00:27:26 let's alone what to say. And so you have to really retrain it and also build these guardrails, otherwise things get out of control. That's so fascinating. And I love hearing about the different AI models that are being built in various industries because there are these very specific nuances
Starting point is 00:27:42 in a courtroom, for example, that you have to consider. But let's think about how things move on from here because you mentioned, for example, that you think that this is somewhat of an ice age, like it's going to happen inevitably. But what does that mean really? Does that mean that, for example, an AI needs to pass its own version of the bar and then I can represent people?
Starting point is 00:28:02 Do you think that it'll just take another generation of people who are maybe more inclined to welcome technology into this space? Like, what will it take for this technology to really be maybe more so welcomed? Despite the legal industry being very protectionist, they have a serious problem, which is that over 80% of people who need legal help can't afford it. And so these courts will have to eventually modernize and allow people to the AI assistance, helping them in all sorts of cases. My role of do not pay is starting with consumer rights. And we have literally hundreds more products to build with AI. And that's what
Starting point is 00:28:37 we're focusing on at first. But I think people will have assistance in courtrooms. I think that consumer rights, they'll have AI general counsel sitting on their shoulder, making sure that these companies aren't ripping them off. And it will just create a more efficient world where companies with business models of kind of ripping off consumers won't be able to continue because the AI will just spam them with lawsuits and legal demands. Yeah. Do you also think there's going to be a deflationary aspect of this? You mentioned that different lawyers are already using technologies like ChatGPT to kind of get their work to the 80% level and then mastering it further. Do you think that
Starting point is 00:29:17 naturally there will be different legal firms that take advantage of these technologies and then can bring down their prices, and then that will deflate the cost of legal services at large? Well, it gets even better than that. There are varying estimates, but some estimates say that customer service costs make up 12% of total corporate spending. If that goes from 12% to 1%, because United Airlines have an AI that handles all of their customer service, there will be huge deflationary pressures. And that's really exciting. And that's how you can see AI lower prices of something that has nothing to do with AI. How does airline tickets have to do with chat GPT? That's the connection. And so it's not just legal services, but also all services could be reduced
Starting point is 00:30:01 just from the customer service angle. I think that paints a very nice picture for consumers. I want to push you on some of the pushback that do not pay as received recently. So I think there's maybe one thread in particular that's gone viral. But I think some pushback is saying these services aren't high quality, but others are just saying, you know, this isn't really using the AI that I thought it was. Do you just have any responses to maybe generally the pushback that the company has received, especially as it's gotten more and more attention in recent weeks and months? Do not pay as in a transition from a template company to an AI company. And you don't need AI to write a angry letter to your former landlord to request a security deposit refund. It's just
Starting point is 00:30:42 a question of the amount and sticking that into an angry letter so that the small claims court can see that you at least tried to ask something. There's not much that can go wrong. What these lawyers have been doing is they've been signing up creating fake cases and creating these dramatic scenarios. So what they'll do is they'll say, I had 100,000 security deposit. Why hasn't their supposed AI told me that the small claims limit is $10,000? And the answer is because that's not a real case. And if you create a fake case, of course you're going to have a fake outcome. And so there's nothing that can go wrong with these consumer rights cases that they've been trying. The true AI is coming, and that's what we've been working on.
Starting point is 00:31:23 Chat GPT is a brand new technology released in November of last year. So I would just say to them, relax, it's not all that serious. It's not the end of the world to get an in-flight Wi-Fi refund. I think that's fair, but I think they would also say something along the lines of what I mentioned before, which is just in some cases, it can be worse to get mediocre legal counsel versus any at all. And so is your goal here with Do Not Pay Currently, while the technology is in the place that it is, to just focus on these consumer rights issues where getting it wrong actually doesn't really have much downside. Or how do you think about maybe the progression? You mentioned ideally you 10x the value that do not pay offers to consumers. How do you think
Starting point is 00:32:03 about taking it to that next level? Yeah, so we're trying to focus on consumer rights, but consumer rights, to your point, could be a $20,000 medical bill refund. And that creates real value for people. My goal is to make it so the average person doesn't need to see a lawyer. If you're being accused of burning down someone's house, you should probably still see a lawyer. But normal people don't deal with that. They just deal with Comcast, United, their hospital, etc. And that's what we're trying to automate with Do Not Pay. The second thing is that we really want to test these AI products, much more so when they deal with serious issues
Starting point is 00:32:38 and also when it's not deterministic technology. So going back to the speeding ticket, one member of the Do Not Pay team actually has a speeding ticket case coming up. And we're thinking that we can get around all of these unauthorized practices of law rules if they're representing themselves using technology they've created. And so we're looking to push forward, be very responsible by testing everything before releasing it. And I would say that consumer rights is not rocket science. And so far they haven't actually found where something's gone wrong. It's been these lawyers coming up with theoretical problems, but people really need help
Starting point is 00:33:13 and technology can help them. I want to ask maybe a far-fetched question, but just, if we look super far ahead, how do you see this new technology, AI, in particular, reshaping the legal landscape? And as an example of this, when might a robot lawyer actually write law that governs the implementation of AI within law? Like, that seems like it hasn't happened yet, but something that maybe is inevitable. I think in the next year, it's already secretly writing laws. A lot of laws originate from an overworked staffer in Congress. And I'm sure it's 2 a.m. They need to finish it and they're probably typing in chat GPT asking it to help them a bit.
Starting point is 00:33:54 And so I'm sure in the next year that's going to happen. In the future, I think a lot more court systems will be automated and tools like Do Not Pay will help people. So I think it's a very exciting time. Speaking of an exciting time, I feel like Do Not Pay has done an excellent job of kind of keeping up with the times. And so something I saw you build recently is the chat GPT extension, which you can basically understand the terms and conditions of a website more effectively. So how are you thinking about the gaps that still exist for consumers and maybe what other opportunities there are to give consumers their rights back? When new technology comes out, we think how can it help people?
Starting point is 00:34:33 In terms of conditions is a great example because people can't understand the changes being made, what they're giving up. So this is actually our first public product relating to true AI, where you can paste in any terms of conditions. upload a document, including leases, and it will go through and tell you all the things that are non-standard and also things you should watch out for. And that's really important because people don't have time to read a hundred-page document, which is what are some of these leases are, especially in New York. And so that can be really helpful. Really, we're trying to be like
Starting point is 00:35:05 David and David versus Goliath, so the AI can turn David into Goliath versus Goliath and help ordinary people. Something that's coming up in my mind as you're speaking is a question around why the law is so complex. And maybe this is such a silly question because there's nuance to life and therefore there's nuance to law. But why are so many laws 100 pages long that no regular consumer can actually read through and understand? Like should it not just be a rule system that you could even effectuate through code, for example? I think there's a lot of lobbying that goes on. The big companies spend huge amounts of money shutting consumers out of their rights. And every year it seems to get worse. There are lots of positive laws that have come out like the CCPA and other laws
Starting point is 00:35:53 around that. But I think big companies have a lot of influence in creating the laws. And one area that I'm worried about is anti-AI laws. So one could imagine a law that says that no bots are allowed unless they say they're a bot. And you're already seeing companies implement these policies. And that could easily be a law. I could imagine members of Congress signing up for that idea. And that would hurt a lot of consumer rights and automated systems helping people. What makes you think that people will actually implement a law like that? Or are there other rules that you see on the horizon based on your experiences with companies, with lawyers who are pushing back against some of this new technology? The fear when any technology comes out is huge.
Starting point is 00:36:35 And it doesn't help that people like me are staring things up. And fear creates action on behalf of governments and big companies. And so that's very worrying. The good news is that there are a lot of things in the law. We have over 100 years of law to deal with that if you jump through the right hoop and follow the correct processes, you can get results. Like one thing we automate is credit report disputes. And they say that if you submit a correct letter and it's signed in the correct way and all of this stuff, you can dispute something on a credit report. And this allows us to kind of get old or inaccurate items off people's reports. And even if it's coming from AI and we have to tell the big company that this was written by
Starting point is 00:37:17 AI, it's not going to be influenced. They still have to respond to that letter. And so there's all these rights that people have and I don't think that they can strip them all away with one law, but it is worrying about where kind of Congress is heading. Yeah. I love that you said or you mentioned that you like stirring the pot. I want to ask you specifically about within the experiment, that you were trying to enter the physical courtroom with your robot lawyer, you also were pushing to enter specifically the Supreme Court
Starting point is 00:37:48 and you tweeted about offering someone a million dollars who happened to have an upcoming court case that was willing to wear AirPods to kind of circumvent that idea of bringing the AI into the courtroom through that audio technology. What made you want to push for that? I think it would show everyone that AI is nothing to be scared of and it's actually pretty good at understanding lots of documents.
Starting point is 00:38:12 AI even passed the bar exam or a portion of the bar exam. And so it is better than some lawyers. And if it was brought in the Supreme Court and all of the little cases we were dealing with like traffic tickets and consumer rights disputes, people would think that it could help them on the smaller stuff. I didn't expect any lawyers to accept our offer. We got lower court lawyers except our offer, but no one at the Supreme Court. But it was a serious offer.
Starting point is 00:38:37 and if you know anyone, please let me know. Well, I'm curious to hear what you're hearing from the lawyers who are supportive because we've kind of talked about the pushback from the legal world that is not in support of some of these ideas and some of this technology. But it does sound like there are people on board and also people who are using some of this technology today. So what are you hearing from that side of the spectrum? Most lawyers are supportive because they're arguing serious cases
Starting point is 00:39:03 around human rights and other issues that don't really impact consumer. rights. And so there's a judge I talk to in San Diego called Von Campos. We talk a lot about cases where people can't represent themselves and they don't have an attorney and how technology can help them. And so just talking to a lot of the good lawyers gives me hope that things can change. Yeah, I'm very excited to see how this progresses, especially because the thing that's so wonderful about technology is that it improves across the board. So as we get more court cases, as these models improve, they're not improving on a single case basis like a lawyer. They go, they represent someone, they learn something.
Starting point is 00:39:43 It's ingesting all of this information and improving with every case that is fed into the model. So I guess as we look towards, you know, not just one year from now, but maybe five, 10, 15 years from now, anything else you'd like to add about where the future of law might go and how technology may be involved in that future? Yeah, I think there'll be instant class actions. So right now the way class actions work is imagine a big cryptocurrency exchange, steals all the money. Imagine.
Starting point is 00:40:13 It's happened many times. Years of bankruptcy and class actions drag it out. And then almost 10 years later, someone gets a check for like $20. And that's what happens with any wrongdoing, the only people that win are the lawyers. In the future, I imagine a world where you didn't even know that something happened. robot lawyers were fighting to you in the background, and almost immediately they get you the re-finaled something back. And I think it will create a more just world
Starting point is 00:40:40 because the bad people know that if they do something wrong, it might take them 10 years, if ever, that they'll have to pay up. But with enough kind of pressure and technology, it can create a world where consumers fight back almost instantly. So I'm excited about that. You know, one example that I loved from your social media was people who can get money from spam calls. I feel like that's just such a great exemplar of all of us get these goddamn spam calls all the time.
Starting point is 00:41:09 They're really frustrating, and you can't really do much about them, or at least up until now. But you've set up a product where you can actually not just stop them, but potentially make money off of them. Yeah, so there's an obscure 1991 law called the Telephone Consumer Protection Act that allows you to sue these spam callers for $1,500 a call. if you're on the federal do not call list. So what the product will do is it will sign you up to the do not call list with a bot if you're not already on it. And then the next time you get a call,
Starting point is 00:41:40 you can pick up and to the point of using call technology to benefit consumers, these spam callers, they hide behind fake names and fake businesses and they don't tell you who they are. But when they try and sell you something, we've built a trap. And what the trap does is it's a credit card
Starting point is 00:41:57 and you can give them the card number. And when they try and charge you for something, it gets their name, phone number, business address, business name, and uses all of those details to sue them for the $1,500. And I actually got a call this morning that I'm doing the case for. And we at users who is basically their full-time job suing spam callers every day. One user even paid it for a new roof for their house with the money they made from suing these callers. So just things like that can be easily automated with technology. And it's not just AI with FinTech. If you have these credit cards that can get all their information,
Starting point is 00:42:33 whatever it takes will use it to help people. That's great. And yes, I think what we mentioned several times here is important. AI doesn't necessarily need to be the technological solution. But on this idea of technology, why haven't there really been many companies who have developed these systems to fight back for the consumer? For example, I know there are technology companies
Starting point is 00:42:56 that will help you right away. or like help manage your estate planning, things like that. But it doesn't really feel like there's been a company that has tackled this with the scope also that Do Not Pay is addressing. And so has there been a fundamental technology shift or what is enabling Do Not Pay to really tackle this in the way that it is? When I started Do Not Pay, Andreson Horowitz was my first pre-seed investor and I made a deal with them,
Starting point is 00:43:21 which is I don't have to drop out of school. I can stay at Stanford and they would invest. and I had other offers, but I decided to go with them because they gave me that deal. And that allowed me to really take time to develop the right strategy. And with any idea, it seems obvious that something like Do Not Pay should exist, why haven't people done it before and why aren't people doing it now? And the reason that I developed by taking it slightly slower at the beginning was that Do Not Pay is only successful because it's a horizontal company.
Starting point is 00:43:51 The average person, unless you're a bad driver like me, only gets a parking ticket once a year. and so you can't build a real company around that. And so because we've gone horizontal and done all of these use cases over time, it's allowed us to build a retentive customer base, which makes a successful company. But that's taken seven years.
Starting point is 00:44:09 And I see Do Not Pay as my life's works. I think it was this unique insight about being horizontal, plus the effort of seven years to actually kind of build it out that has allowed us to be unique. And there are companies, even offered by lawyers, that allow you to get out of your tickets, but there's no one trying to tackle all of consumer rights with the tools necessary
Starting point is 00:44:28 where we're not even building individual products. We're building tools to build products so that when something bad happens, in two weeks we can come out with something to help people. Yeah, the product cycle has been incredible. And as an SEO nerd, I also dug into the different content that you've produced. And it's so wide ranging. And I love the idea that actually when you are that horizontal company, you can do keyword research, for example,
Starting point is 00:44:55 just to find out all the consumer search volume for issues they're facing and you can somewhat use that data to figure out, okay, this is the next feature or the next product that we're going to build versus if you are constrained in a very specific area or use case, you can't really expand in those ways or use that data to expand.
Starting point is 00:45:14 So I think that's great. Josh, anything else you want to leave listeners with about your company or the recent experiments that you've been running and how those have gone. We have some really exciting announcements coming, including around in-courtroom usage of AI. Awesome. Well, I want to end off on just a fun little game because I know we've been talking a lot about the law. And I did some research and found some very strange laws from around the country. And I wanted to just probe you with these and see if you could guess accurately if they were
Starting point is 00:45:46 in fact true or false, if they are actual laws that exist. Oh, no. When I don't do well with the game, the lawyers are going to go crazy. No, they're going to use this as a proof point that you are not fit to reshape the legal system. But there are truly obscure laws. So the first one is true or false in Arizona. Donkeys cannot sleep in bathtubs. I'm going to say that's true.
Starting point is 00:46:12 You're right. And the law was enacted in 1924. So that one is indeed true. Okay, the next one is in Rhode Island. coins cannot be placed in ears. I'm going to say that's false. You got that right again. But in fact, that is a law in Hawaii,
Starting point is 00:46:32 which was enacted in 1847. So an old law. I didn't guess that. I just knew it. Definitely. Okay. Next one is in Alabama. You cannot have ice cream in your back pocket.
Starting point is 00:46:45 I'm going to say that's true. So that one's false because it's not. Alabama, but it is, I believe, a law in Kentucky from the 1800s. You know, something as I did this research that was fascinating is just to see how many laws still exist from hundreds of years ago. And I wonder actually if you know much about this, Josh, is there a process to like expire laws that no longer apply? No, there's not. And the laws you mentioned probably aren't enforced, but some of really obscure laws are still enforced. Like in Utah, hotels can't give alcohol for free because of some religious laws that have existed for a while. And so all of these obscure
Starting point is 00:47:25 laws, that's what makes America great. Every state has all of these different laws. Exactly. We'll do one more, which is in New Jersey. Customers are not allowed to pump their own gas. Oh, that's true by a mile. Yeah, that's true. That's true. And that was enacted in 1949. So I believe it's one of, if not the only state where that's still true. But you know, you're not covering some of those laws, but I think I would really encourage people to go to Do Not Pay and look at the super wide range of things that you are supporting consumers with because it's really eye-opening.
Starting point is 00:47:56 Yeah, thank you. All right, Josh, thank you so much for joining us today. Thanks for listening to the A16Z podcast. If you like this episode, don't forget to subscribe, leave a review, or tell a friend. We also recently launched on YouTube at YouTube.com slash A16Z underscore video, where you'll find exclusive video content.
Starting point is 00:48:17 We'll see you next time.

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.