Fastcase’s Artificial Intelligence Sandbox

Episode Notes

Transcript

Digital Detectives

Fastcase’s Artificial Intelligence Sandbox

04/23/2018

Intro: Welcome to Digital Detectives, reports from the battlefront. We will discuss computer forensics, electronic discovery and information security issues and what’s really happening in the trenches; not theory, but practical information that you can use in your law practice, right here on the Legal Talk Network.

[Music]

Sharon D. Nelson: Welcome to the 90th edition of Digital Detectives. We are glad to have you with us. I am Sharon Nelson, President of Sensei Enterprises, a digital forensics cybersecurity and information technology firm in Fairfax, Virginia.

John W. Simek: And I am John Simek, Vice President of Sensei Enterprises. Today on Digital Detectives our topic is, ‘Fastcase’s Artificial Intelligence Sandbox’.

Sharon D. Nelson: Before we get started, I would like to thank our sponsors. We would like to thank our sponsor SiteLock, the global leader in website security solutions. Learn more at  HYPERLINK “http://www.sitelock.com/legal/digitaldetectives”sitelock.com/legal/digitaldetectives.

We’d also like to thank our sponsor PInow.com. If you need a private investigator you can trust, visit  HYPERLINK “http://www.pinow.com/”pinow.com to learn more.

John W. Simek: Our guest today is Ed Walters, the CEO and co-founder of Fastcase. Under Ed’s leadership Fastcase has grown to one of the world’s largest legal publishers currently serving more than 800,000 subscribers from around the world.

Before founding Fastcase, Ed worked at Covington & Burling in Washington D.C. and Brussels. He worked in the White House from 1991-1993, first in the Office of Media Affairs and then in the Office of Presidential Speechwriting. He is an adjunct professor at the Georgetown University Law Center where he teaches the Law of Robots, a class about the frontiers of law and technology.

It’s great to have you with us today, Ed.

Ed Walters: It’s great to be here, John. It’s funny. When Fastcase was just getting started, I remember meeting you and Sharon at the Virginia State Bar Annual Meeting and we had like the crummiest, smallest table. We were like a brand-new exhibitor and you guys were rock-stars, man, like everybody knew John and Sharon, and you were so nice to us, like helping us break in and telling us what to watch out for and what not to do, and it’s great to kind of bring it full circle to be on Digital Detectives with you all today.

Sharon D. Nelson: That’s nice. You’ve come a long way, baby.

Ed Walters: We all have.

Sharon D. Nelson: No more rickety tables and you’re certainly no rookie in the business anymore, that’s for sure. We were very excited when we talked to you a few months ago and you were telling us about Fastcase’s new initiative called the AI Sandbox.

So, tell us a little bit about what that is and how you came up with the idea, because I am sure, well, John and I know something about it, our listeners probably do not.

Ed Walters: So, the idea is that lawyers, law firms, information professionals in firms should be using AI for themselves, right? So, in finance, in sports, in marketing, in all kinds of other business, people use Artificial Intelligence to gain insights, to get some like advantage over their competitors almost everywhere except law.

And so, there’s a lot of people talking about AI right now but what they really mean is firms should subscribe to some end-to-end tool that uses AI at some point in the process; and so, we thought it will be better if law firms and lawyers could become makers if they could learn to use the tools themselves and to create things anew.

But it’s hard to do, right? In on-premises servers you can make a mistake, you don’t ever want to give your data to some third-party vendor, and so, the idea was to create kind of an environment for experimentation, software developers often call them sandboxes where people can mash up their legal data with third-party legal data, use AI tools to analyze it, see what works for them, see what kind of insights come out, maybe even make some mistakes, try some things that don’t work, but in an environment that they control and that they get to experiment with.

So, every business in America has some sort of R&D budget, but law firms never do. Law firms should do research and development. They should do it just like any other business does and so they should have a place where they are doing some experimentation, pushing the frontier, and trying to figure out what’s possible. So, that was the big idea behind the AI Sandbox.

John W. Simek: Well, Ed, Fastcase has a lot of partners providing tools and data, can you tell us who you’re working with?

Ed Walters: Sure. So, the big idea is that we’ll be kind of a meeting place, right? We will combine law firms’ own data, they will control their own Sandbox, no one else gets a look at it, and then we’ll bring data in from Fastcase and from other partners.

(00:05:04)

We just acquired Docket Alarm, the major kind of Docket Analytics company, and so we have data from Docket Alarm. We have this skunkworks project inside of Fastcase. It was really just our own R&D experiment to try and download the metadata from PACER, just as kind of like a — kind of a wild hair day we decided let’s see if we can download all of the metadata from every case file in PACER for like a whole year.

And so, we did it, like 750,000 rows of metadata from PACER and so we’ve put that into the Sandbox as well. Now, it’s — I think it’s more than 2.5 million cases worth of data and then we have all these partners, who are also putting data in.

The most prominent of them is IBM Watson who has put a bunch of their AI — APIs (Application Programming Interfaces) into the Sandbox. We work with LexPredict and Contract Standards. There’s literally dozens of partners who are putting expert witness data or legal analytics data or contracts data into that Sandbox and then a whole lot of AI tools providers who are also putting their data in there. You can get a whole kind of roster of them if you want at  HYPERLINK “http://www.fastcase.com/sandbox” fastcase.com/sandbox, but it’s really cool to see.

So, now we have about 10 firms who have like kind of created their own Sandbox space with the AI Sandbox and they’re creating all kinds of crazy projects, they’re great.

Sharon D. Nelson: Well, it sounds like you’ve got some really interesting partners going here. I’m curious though for the projects that the first law firms are working on, what kind of projects are they working on and what AI tools are they specifically using, Ed?

Ed Walters: So, one of the first ones is firms want to understand their own pricing, how much they charge for clients for a certain matter, and maybe that’s obvious. Clients more than ever, especially corporate clients are saying they want to move to fixed price engagements or alternative fee agreements. And in a situation like that, law firms just have to know how much those engagements cost them, but it’s really hard to know, right? And so, that’s one of the first projects we’re seeing in firms, they’re trying to understand their own costs for certain kinds of legal services.

Another one is there’s a lot of firms that are trying to understand expert witnesses better, so they have a roster of expert witnesses who have testified either on behalf of their clients or adversely to them in litigation, and so firms want to have more information about expert witnesses, when they were challenged, how they were challenged, whether they survived the challenge, whether they were successful, whether clients believe them?

So, expert witnesses are a good example of partner data, and we have a number of partners who are providing that kind of expert witness data and then law firms are taking their own expert witness data and mashing it up together, and then using tools to understand and visualize that data. That’s the dataset.

I would say IBM Watson tools are very popular here. They’re also using some tools that are open-source. So, one that they’re using is called Kibana, it’s a tool of Elasticsearch that allows people to visualize datasets, but over-time I think we’re going to see a lot of different both datasets like maybe census data or PACER data, judicial opinions or statutes, and then a lot of the tools from IBM Watson. We think of Watson as the kind of big blue obelisk that beat Ken Jennings in jeopardy, it’s not, it’s like 62 different pieces of software that just solve a particular problem. And those APIs were used in combination to win at jeopardy but you can combine them in a bunch of other different ways for legal work.

I’ll give you one other example. There are APIs in IBM Watson that do personality insights. So, if you go into a big-box store, a salesperson within five seconds of you walking in the door is going to look at your shoes and your belt and your watch and your handbag and your glasses, and they’re going to have you put into a category of buyer right-away, and they’re going to use that persona as a way to understand who you are and decide how to sell to you. So, lots of businesses do these kinds of personas.

Lawyers for the most part don’t, but so, I think larger law firms are starting to do this. They’re starting to say, we can use these AI insights to understand judges better, to put them into a persona, to put opposing counsel into like kind of a persona, to put jurors into these kind of templates for persuasion.

(00:10:05)

Not to say that we know everything about them, it just helps you understand who they are and how to reach them better.

John W. Simek: Just as Zuckerberg, you’ll find out more.

Ed Walters: The ultimate big data application if we could just harness the power of Facebook.

John W. Simek: Yeah, I think you better — maybe should start looking for another partner in. So, Ed, you talked about Fastcase data in the Sandbox, can you kind of expand on that a little bit more?

Ed Walters: Yeah, so, we are sitting on one of the biggest legal datasets in the world. There’s 255 million documents in the Fastcase and Docket Alarm databases, and so, we have a ton of information about judges or law firms or clients, and so armed with the kind of whole legal database of judicial opinions and statutes, all of the motions and pleadings from the Docket Alarm database and then all of the metadata about filings in PACER, what the pace of things were, when they happened. It’s an amazing amount of data, and so we’re putting some of that data into the Sandbox as well.

This was one of the impetus moments for the AI Sandbox. We were talking to partners who said, look, we would love to cross-tabulate what we know from our files with what the judicial opinion database knows about cases in the Virginia courts, about litigation in the Second Circuit.

But, everyone we talked to about this, all of the kind of big enterprise players here want our firm to wrap up all of our data and send it to them so they can do the analysis and then send the results back to us. And there’s no law firm in America that’s going to do that, that’s crazy town. I mean, can you imagine a lawyer being so careless with client data?

Sharon D. Nelson: Yes; yes, we can. We see that all the time.

John W. Simek: I think it’s been done before.

Ed Walters: Perfect. But so, we said, look, there’s no reason in the world why you should do that. If we create a Sandbox environment that the law firm controls and they only control, we can give them the data. They can do the analysis themselves, and five years ago, even three years ago, that was just crazy. No one could work with these AI tools because they were too magical.

But now, they’re more democratized than ever. It’s easier to use them. If you, with some very lightweight training, that’s a part of the AI Sandbox project, anybody can use these tools to do the analysis. And I say this as someone who is like a government major and philosophy minor, I got a D in calculus for God’s sake, I’m no wizard, I’m a muggle, but, the tools are more democratized than ever. So, even a non-scientist, like me, can use these datasets.

Sharon D. Nelson: Well, that kind of sounds exciting and I hope you’ll do a video or something that everybody can see because they’re afraid of the Sandbox because they do think it would take a lot of training. So, maybe a public video would be good.

Ed Walters: That’s a great idea. And we are going to try to publicize the success stories. So, at this summer is American Association of Law Libraries’ Annual Meeting. We will have a panel of people who build things in the Sandbox, where they just show their work.

We just did one at the Ark Group’s Knowledge Management Conference in New York City, where we did the exact same thing, and what I really hope to do here is not showcase how awesome Fastcase is, although I always love that. I want to showcase how awesome the projects are the people build in the Sandbox.

I want people to show their work and say, wow, look, I took this big mess of data, maybe just all the documents in our firm is like databases and organized it, and pulled out who the lawyer writing it was, who the judge they were appearing before was, and then just visualized it. I mean, that alone would be such an amazing tool, it’s hard to do today but pretty easy to do with AI tools.

Sharon D. Nelson: Well, just make sure you get it up online because that’s really the democratization in part will be spurred by being able to see it online as opposed to seeing it at a conference, if you know what I mean.

Ed Walters: No, I totally know.

Sharon D. Nelson: It totally, totally spreads faster. So, let’s talk a minute about legal AI, and I hate to phrase it this way because the words “me too” have such another meeting these days, but is the AI Sandbox just another Me Too AI project?

Ed Walters: Yeah, I hope not. I mean, I really worry about the kind of peaked expectations around AI where we are in Gartner’s Hype Cycle and there’s a lot of people who are talking about Artificial Intelligence right now.

(00:15:05)

It seems like it’s mandatory for legal-tech conferences or just legal conferences in general, you have to have a panel about AI, and we’re starting to create like a talking class about Artificial Intelligence and there’s a lot of talk about it, but there’s not a whole lot of work in AI.

And I just would like to see some of the demystification of it, instead of the AI Sandbox being yet another tool that uses AI that you sort of apply in an end-to-end way. I’d like to see people really building with it and seeing that there’s nothing magical about it. It’s not unicorns and pixie dust, they’re just tools. It’s like a pivot table in Excel or a regression analysis.

It will show you interesting things, but not by magic, just by math. And so, my hope is that we can move from an era of read-only Artificial Intelligence or people might subscribe to a tool that uses Artificial Intelligence. Think of like electronic discovery, where you apply the tool in a simple way but you never really get under the hood. Move from an era of read-only Artificial Intelligence to read/write Artificial Intelligence where people use the tools and build their own AI applications.

In the old days, typesetting was really hard to do. It required some fancy software and some real knowhow to put together a newsletter or a properly formatted article or something, but now, anyone can do it, that’s just desktop publishing software. You can do most of that stuff in Microsoft Word. Because the tools were democratized and were in that world right now for Artificial Intelligence, the raw materials and the tools that make up even complex services, like IBM Watson, are now easier to use by non-technical users, which is awesome, that’s really great.

And the one thing I want to call out here is that this will work to the benefit of clients. We lose sight of this a little bit in the conversations about AI magicalism but all of this should go to the benefit of our clients. We should serve them better, we should get better answers for them. Clients ask us, lawyers, these kind of bet-the-company questions, these make-or-break questions, and sometimes we answer them with hunches, based on five matters we’ve seen before, we think it looks like this.

So, when they say, should I take this settlement offer? Do we really have the data to give it an informed answer? I don’t know, but the hope is, with AI, we can look both in our databases and other databases to really give an informed answer, to give better educated answers, not with conjecture but with data.

John W. Simek: Well, Ed, what’s all the fascination with AI in law right now? You kind of touched a little bit on that but why now?

Ed Walters: Well, I feel like maybe the big driver for it is Access to Justice. I hope that’s what the big driver is for it. There is a very famous 2014 American Bar Foundation survey that says that 80% of people who need legal help, aren’t getting it. And in some cases, it’s because they don’t feel like they can afford it, but that’s not the biggest reason.

The biggest reason is, people don’t know that they need legal help, and so they are going without it where they’re kind of going it on their own. And, if you look at any of the studies of it, people who go it on their own lose the vast majority of the time. Navigating the legal system without a guide is a fool’s errand. I mean, you will lose 95% of the time. And in many cases, when you’re pursuing the same matter with a lawyer, you’ll win 95% of the time. But what does it say that we’re only serving like 20% of the market.

So, I think that there’s a big push right now in lawyers, and I would say also like paralegals and non-lawyer professionals who are trying to help people and bridge that Access to Justice gap. It’s certainly my opinion. I think the opinion of many other people that we can’t do it with pro bono work, there’s not enough pro bono hours out there to do it. But AI might help if we can build software at scale that can help people, not like a Watson jeopardy machine maybe, maybe like a neurologic-powered decision tree that is created by experts or maybe lightweight AI that can answer frequently asked questions or help people to fill out forms better with Document Automation, then we can use AI and software as a force multiplier for Access to Justice.

(00:20:00)

So I hope that’s what’s driving it. It’s certainly exciting. I mean artificial intelligence holds the promise that we can give better advice, that we can give it at scale, we can give it at lower prices, we can deliver better legal services. But again, I really hope that access to justice is what’s driving it.

Sharon D. Nelson: Well, I know you do, and I know how sincere you are about it, but I guess I am cynical and although I hear a lot of talk about it, and some people are very genuine and sincere, there are a lot of people in it flatly for the money and the profit. And the access to justice for many vendors particularly seems to be kind of a song and dance that they do.

So I hope you are right too, I would love to be wrong here, but I am not sure I am.

Ed Walters: There is certainly a lot of money chasing it. I mean you are not wrong about that.

Sharon D. Nelson: Yeah, there’s a lot of money chasing AI for sure, for sure.

Ed Walters: There are a lot of venture funds and others who are very interested in anything that — this month it’s artificial intelligence; next month it will be blockchain. There is always that kind of flash bang, of the moment buzzword that will attract attention and press and money.

But why now? This stuff has been around since the 1950s. It has had big demonstration projects before. The biggest most recent one, the victory over Brad Rutter and Ken Jennings in Jeopardy was 2011, seven years ago.

Sharon D. Nelson: Right. Right. Well, its held on to some of its gloss I think because it has gotten a little bit more practical and we have seen some demonstration of success.

But I think it’s probably time to take a break, isn’t it John?

John W. Simek: Yeah. Before we move on to our next segment, let’s take a quick commercial break.

[Music]

Sharon D. Nelson: At least 80 of the 100 biggest law firms in the country have been hacked since 2011. Protect your firm and your clients from cyber attacks with SiteLock. Their industry leading cloud-based suite of website security solutions includes website scanning, web application firewall, including DDoS mitigation, and 24×7, 365 US-based customer support. Give your firm and your clients’ peace of mind knowing their information is secure. Learn more at  HYPERLINK “http://www.sitelock.com/legal/digitaldetectives”sitelock.com/legal/digitaldetectives.

[Music]

Advertiser: Does your law firm need an investigator for a background check, civil investigation or other type of investigation? PInow.com is a one of a kind resource for locating investigators anywhere in the US and worldwide. The professionals listed on PInow understand the legal constraints of an investigation, are up-to-date on the latest technology, and have extensive experience in many types of investigation, including workers’ compensation and surveillance. Find a prescreened private investigator today. Visit  HYPERLINK “http://www.pinow.com/”www.pinow.com.

[Music]

Sharon D. Nelson: Welcome back to Digital Detectives on the Legal Talk Network. Today our topic is Fastcase’s Artificial Intelligence Sandbox. Our guest today is Ed Walters, the CEO and Co-Founder of Fastcase.

John W. Simek: Well Ed, what’s going to come next for legal AI? In other words, bring your crystal ball out there and tell us where we are going to go from here.

Ed Walters: Well, I think it’s makers. I think this is sort of why we bring out the AI Sandbox when we do, and maybe it just seems self-serving, but I really do think that is the next trend. I think that people will stop being consumers of AI and start being makers of AI. And I am not talking about data scientists; I am talking about lawyers who are curious, lawyers who want to know more.

If you have a bunch of lawyers who are all using the same tool, they are all using Docket Alarm or Recommind or some other tool that has artificial intelligence baked into it, they are all going to reach the same insights. There is no differentiation there.

In finance, hedge funds use artificial intelligence to get just a little more information and just a little faster, and that alpha created one trillion dollars of value in finance. What are we going to do in law?

So I think that the read-only AI is the new floor. If you are not doing that, you are missing out. You have to do at least that. You have to have at least the analytics baked into third party tools. But the ceiling is going to be defined by what people make themselves with those AI tools.

(00:24:50)

And we were talking before about why now, I will just say there’s more data now, there’s more legal data available than there ever has been, and the tools are democratized now. They haven’t always been. It used to cost like $300,000 to start using tools like IBM Watson; now you can start doing it for free, and you can pay a quarter of a cent per API call, so they are more democratized than ever and the methods are better than ever.

So more data, more democratized tools and better tools create this renaissance for AI makers, I think that’s what’s next.

Sharon D. Nelson: People talk a lot about insights from artificial intelligence, but what they really are talking about I think is data. Isn’t legal data an important precondition?

Ed Walters: It absolutely is, and people often talk about garbage in, garbage out, but with artificial intelligence, and especially machine learning, you can only train it with data.

And so it’s interesting, a lot of the early AI Sandbox projects are going to be just cleaning data. They are going to take like big masses of unstructured data and try to pull out information like how much the matter was worth, or who dealt with it or what stage in litigation it was. Data is going to be what powers this makers revolution and the next phase of AI.

In fact, I have a book coming out about this in the late summer called, ‘Data-Driven Law’, a collection of essays from luminaries like Kingsley Martin or Stephen Wolfram, a lot of real pioneers here. So it will be a good case for this where we say — like Ken Grady and others say, data is going to be the precondition for delivering good legal services in the future, instead of hunch-driven law, data-driven law.

John W. Simek: Well Ed, the book you referenced, is that the one that you are currently working on or something else?

Ed Walters: It’s done. No, it’s in production. It’s coming out from Taylor & Francis, I am told probably in October, but it’s possible they are going to finish it a little bit early, so it’s going to be a lot of fun.

John W. Simek: So you have no more publishing projects on the horizon for you?

Ed Walters: I didn’t say that.

Sharon D. Nelson: Well, you have got a new publishing brand, right?

Ed Walters: We do. So we have just started Full Court Press here at Fastcase. It’s going to be our publishing imprint. It is the imprint that’s publishing RAIL: The Journal of Robotics, Artificial Intelligence & Law, which is new from Full Court Press.

And I can’t remember if I told you all the story of this, but the name comes from a Malcolm Gladwell story about how David beats Goliath; may e time for another podcast.

Sharon D. Nelson: Yeah, I know that story, I know that story, but I don’t remember Full Court Press in the story.

Ed Walters: Yes. So the basketball coach had the girls’ basketball team that was not going to be taller or more athletic or more experienced in basketball than the kids they were playing, and he knew he couldn’t teach them to shoot better or rebound better, but he could teach the girls to do a full-court press.

And so this scrappy girls’ basketball team who wasn’t beating their opponents like 65 to nothing, but they were beating them like 6 to 4, because the other team couldn’t rebound the ball.

And so the big idea is if you want to beat Goliath, don’t play Goliath’s game. You have to do things differently and full-court press is an example of that. When it came time to name our imprint, Full Court Press seemed like a natural.

Sharon D. Nelson: It does and it sounds exactly like you guys. So I think that’s wonderful. I love the story behind it too.

Thank you for joining us today, Ed. I know you are a busy guy and you always take time out for us and I am glad we go way back and we remember our humble beginnings, all of us, and it was kind of great back in the old days. It was simpler. Life was simpler then. But we really love having you on the show, and thanks for telling us about the Sandbox and I hope folks will take a look at it.

Ed Walters: I am so grateful. Thank you for having me on.

John W. Simek: That does it for this edition of Digital Detectives. And remember, you can subscribe to all the editions of this podcast at  HYPERLINK “http://www.legaltalknetwork.com/”legaltalknetwork.com or on Apple Podcasts. And if you enjoyed our podcast, please rate us on Apple Podcasts.

Sharon D. Nelson: And you can find out more about Sensei’s digital forensics, technology and cybersecurity services at  HYPERLINK “http://www.senseient.com/”senseient.com.

We will see you next time on Digital Detectives.

[Music]

Outro: Thanks for listening to Digital Detectives on the Legal Talk Network. Check out some of our other podcasts on  HYPERLINK “http://www.legaltalknetwork.com/”legaltalknetwork.com and in iTunes.

[Music]

Continue Reading