 now. Hi everyone, welcome back to the second class of this year's MIT IAP computational law workshop course. But before we get started today, DAS is going to be going over a couple of housekeeping items. So I'm going to pass over the Florida DAS. Great. Thank you so much. And Brian, if you'd be kind enough to call up some of the class slides that would be terrific. Okay, thank you. So, welcome back everybody. I want to say that we had our, I'll just take care of some housekeeping things in a little bit of context setting we had our first office hours last week on Tuesday and I think it was successful people seem to like it. And what we, the rhythm for those of you that weren't part of it is we basically went more deep into the substance of the topics and so people had some nice lines of questions. And we had an opportunity to discuss them to in that vein. So as a speaker, you'll be hearing from the co-teacher, the TAs, some lightning talks later today, and they're going to go even deeper into different what I think are seminal aspects of computational law. And we're going to use the office hours next week to really get into the dialogues. And so, so have a listen today kind of try to digest what you hear. And then do the remarks later with some catalyzing questions and we can get into it more deeply in office hours, and we have the telegram channel at long last up and running. And so everybody's heart it took an entire village to really get that going. And so if you're in it. I encourage you to use that as a back channel during class and to, you know, maybe even start grappling with some of these topics and of course our main speaker. But not yet in it. We have a link that we have sent and I'll send out as part of the follow up where you can request to join pro tip. You need to make a telegram username so that we can invite you into this kind of closed channel, and we have a link to that right in the form. And now I sent I also gave a little bit of a deep preview to class number three, where we've got a couple of great speakers Senator Chris Rothfuss, who's going to talk about, I think groundbreaking digital identity legislation that some of us have been involved here, Brendan Mar, TMA, myself and others. And we've got links to some of that. And, and also we'll hear from New America Foundation on the essence of property ownership in the digital age. So it's not too early to start taking a peek at some of that stuff. But our main course today is standards and that underlies so much of computational law. Without further ado, I'd like to hand it back to our erstwhile MC TMA to set up our first speaker, my friend and colleague Ken Jones so TMA take it away please. Thanks, Brian, my moving this slide over to page. Great. Awesome. So, today, like that said, he joined by Jones. Next page. Mr. Jones is a chief technologist at tandem bomb kill LLP, where he oversees various aspects of technology at his firm in his role as CIO. He leads efforts to support TK is computing environment and infrastructure to ensure that client data in the cloud is processed and protected with highly skilled and respected leading edge business partners in the technology industry. Mr. Jones also leads and supports various TK programs in the areas of security compliance, business continuity and firm administration. Okay, Ken, the floor is yours. Well, thank you so much to me and as it's a true pleasure to be here. And it would be fine to move the slides forward into the agenda for today. As I indicated, the area of standards is is a vitally important area, it lends itself to interoperability communication between companies, and a whole host of other benefits and I will do my utmost to, to try to convey some elements of information and wisdom as it relates to this. The first thing I wanted to say is that, you know, everything in life can be a learning experience as Daza has pointed out. A good element of my job is one serving as the CEO of a law firm standards and things like that are very interesting. There's an element of my job that relates to help desk tickets and broken computers and cloud platform technology and everything in between. So, just to in preparing for this opportunity to speak on this topic, you know, it encouraged me to speak with many experts to do independent research and I myself benefited from it. And so it's been very beneficial for me and I'm grateful the opportunity to be here. And it pertains to my professional life just to introduce myself. As I made the same point I was going to make I have a very practical job. You know I'm just a simple chief technologist from a law firm with three offices here in the United States. My ideal is a litigation boutique, a full service firm serving primarily Fortune 500 companies. You know today, we'll be discussing many theoretical topics topics that are also practical and in practice but other things that are more forward thinking and, and my anticipation and Daza and many others of the industry that we are all working with and studying. I think it's important to note that a balance between practical work and theoretical work is is something we all have in our lives. It's, it's what makes life interesting and it's, you know what, you know, makes us valuable in the marketplace. So today, we'll talk about some standards in business in general will then shift into legal standards and talk about the history of legal standards in the marketplace. We'll talk a little bit about technology fragmentation in legal, and, and finally we'll summarize many of the current and emerging legal standards. So we can go forward thank you very much. And we'll talk to begin with all about standards in industry, I've worked in the legal field for 18 years. Before that I worked for a big pharma company Bristol Myers Squibb for 13 years, pretty much my only two jobs. And at Bristol, at the time they were a large consumer products and pharma company. You know something like Procter & Gamble, and or J&J, you know, and so forth. And so, as we sold to large customers like drug retailers like McKesson companies like Walmart, some others that aren't in business anymore. There was a tremendous needs for standards to just handle day to day business, like taking an order, or making a shipment or invicing, you know, the kinds of things that we all see and do when we say shop online with Amazon or any other number. And back then, some standards bureaus emerged. We had the American National Standards Institute, which was primarily, you know, an American thing. The nation said something similar called ed effect, which was primarily European, not exclusively. And business and enterprise had tremendous gains because instead of building one interface between companies, all of your customers for orders. We had a common one. You know, you could reuse work, you could follow best practices, the standards were set by, you know, organizations, not by say a leading market leader like say Microsoft might set conventions for outlook, not that they do it poorly, but you know, it's still maybe better when you have a cross functional team doing that. And, and so I learned early in my career in the 90s, some of the values of standards in in the world. I think maybe you might want to make a brief additional comment in this area. You know, I, you know, I do have some comments but I think my, I'm liking the flow of the talk so I don't want to, I don't want to interrupt it. If you don't mind. Thanks. So, you know, be it understood that, you know, standards have tremendous value in industry, and will now shift gears rolled to the next slide, and begin to focus and hone in a little bit more on on our industry on the legal industry. So, again, this slide is meant to talk about the past, the past can be boring but we can also learn from the past and from history. So, there are really two primary standards or communication interchanges that they're not confused but used interchangeably even though they're not really interchangeable, but there are uniform task based management system task codes. And then there's this leads standard legal electronic data exchange stand which sends data around from point A to point B. So, essentially what that is in practice is as timekeepers attorneys, you know, or power legals legal assistance work on tasks for clients. They will enter their time into a system I spent two hours working on a legal brief. I spent an hour doing research I spent all day in a trial, you know, and so forth, and that data is entered into time and billing systems and transmitted to our to our clients and so the test codes were designed by the ABA. The most common ones are something called a litigation test code set, which is this five series is five phases, things like, you know, discovery research free trial trial and a ballot. And those codes are used to generate time entries which generate invoices which are transmitted to clients to generate payments. These standards, which were also made with, you know, with the assistance of Pricewaterhouse cupboards, they're widely used, you know, this is the number one way that today law firms bill for time. And, despite some weaknesses that we'll discuss later it's still one of the ways that law firms are compared and contrasted with with others in terms of how efficient, we might be who handles like matters better. To a certain extent, there's efforts to try to use them for budgeting and very, you know, other other types of processes in the legal field. So the key takeaway here is that there is at least one common standard in legal it's it is leads and the test codes, and it's used extensively for billing but in other areas, there are fewer, you know, standards. Now, to what extent is that important, and this is a little bit of a shift in focus but it's important to sort of set the scene and properly analyze the situation. You know, the, the legal technology industry is fairly fragmented to compare and contrast it to other areas, you know, almost all of us, not all of us but almost all of us would use Office 365, you know, for email and Excel and Word and PowerPoint and so forth that's that's very consolidated in the corporate world, ERP systems, enter size resource planning or CRM systems, customer relationship management, they're moderately consolidated sales force, for example, is a customer relationship management system that has a pretty good significance. We've all heard of them. They're a common company, you know, it's, it's fairly common when a sales person for a fortune 500 company might go out to make a call and transmit information back to their company that they're using sales force. SAP is a large enterprise resource planning ERP system. It's, it's fairly common that SAP or something like Oracle financials could be in used in in large institutional clients to handle many functions of an organization. Example would be companies with say order to cash initiatives, this means everything from taking an order to getting a payment would be handled in a single system. You know, other opportunities for this might be companies that have a single chart of accounts. You know, that's, that's your general ledger, you know, and, and having one of those throughout a large institutional company. Well, it makes all the sense in the world, but in practice that it's used to be at least quite rare. And now it's increasingly common. So, you know, the legal industry tends to however have different players and different verticals, you know, companies that do practice management or matter management or document management systems DMS or time and billing. So the most part are different companies, you know, I manage in net documents might be in the DMS space they are in a DMS space, but they're not really in in many other spaces. And so this presents issues, you know, and one of the biggest one is, how do you get data back and forth between these systems. And coming back to 10 minutes ago, you know, EDI standards with orders and financial information and shipments and so forth, made that much easier in general industry. It's still a big challenge in legal. And so, when you begin to factor in looking at the final bullet, the fact that we have entities like Renyan court which is a containerized way to deploy applications and something that's being advocated, fairly only by the largest law firms in the world like am law 50 or 100. And that's a technology where you can render applications to two groups in a private cloud, which is, you know, desirous for the types of firms that work on like M&A and financial information that's, it's exceptionally high security. But you know then you have public cloud entities like like I manage in the cloud, or Salesforce is something like that. Lots of different approaches over in the top platforms maybe better way of putting it over and on top of different companies. So it just offers the legal tech industry. A lot of consensus opportunities opportunities to consolidate but until that happens standards and ways that these companies can, you know, interact together in a in a common manner is exceptionally valuable, even more valuable, perhaps in the rest of the technology world. So, let us now dive into the material and I apologize for taking so long to get here. There's a entity known as the Sally Alliance, and it is generating not for profit legal industry standards. It's a very very very valuable group, a group that's doing great work, and why is that, you know, so one reason is they're working towards a standard called a legal matter standard specification LMSS. They have two versions out. You know the second one is fairly new but this is designed to be a standard that allows every system to use common metadata to classify data to testify matters excuse me, or to more importantly, or to be as importantly to classify common work tasks. It's, you know, the process of taking a deposition, or, you know, attending a court conference or any other legal work task is very hard to budget for it's very hard to price it's very hard to do anything if you don't have a common understanding of what that is. And one of the biggest challenges of that is that it's totally different in different practice areas. A deposition for our products liability litigation matter was very different from one maybe in family court or any banking or lots of other different areas. So one of the core things LMSS looks to do. And, and I spoke with, and they were very gracious with the time Peter Buck from from net documents and Toby Brown from Perkins coey is they look to classify legal work, you know so if you have a document in a system, you would use this specification to understand that it's an IP document. It's a litigation document it's a banking documents it's a merger and acquisition document and so forth. You know, prior to these standards, everybody had their own way of doing it. Probably worked for most people, but virtually impossible to communicate those across systems and you know across different entities law firm to client, you know, and well, you know, corporate legal department to law firm and so forth like that. You know a lot of great work being done there. And one of the reasons it's really good bouncing back to a couple slides ago, not that we need to do that but just to relate back to it is that the leads standards and the task codes were built primarily by the ABA and by press waterhouse Coopers. So, and the people did a great job I'm not being critical but it was created by lawyers, you know, and with that, with their perspective on on the issue, whereas Sally is a broader consortium. They have folks from the courts they have, you know, legal operations specialists they have clients they have offers. It's very hard to get broad representation of stakeholders who can all hash out issues. It tends to I think, probably make progress a little bit more slower and laborious but at the end of the day, it cranks out better work product. It really does offer a great promise because of their, their methodology and instruction they receive funding from all kinds of groups one of them to say the Delta the International Legal Technology Association. You know, so they can remain, you know, standard agnostic and do what they truly believe is best for the industry. The vision to date, you know, and I'm, I'm no expert, like I said, I'm a, I'm a practitioner, more than a, you know, a deep thinker but they've done a lot of work on legal matter specifications, they're looking into billing definitions they're looking into the other types of litigation. But the bottom line is, you know, the hope is that these various taxonomies lead to common understanding of tasks and activities. And, you know, before we leave the slide I mean that's that's really important because as you try to price or budget matters, unless you have that common understanding. It's very difficult to do it, you know, somebody thinks you're doing something it takes a half an hour other people think that task might take two days and that sounds ridiculous but it's not untrue. You know, there's a really good piece by Jay on legal evolution this whole week, and she's a pricing expert with a law firm and you know I'd recommend checking that out she has a lot of good ideas but a lot of the things in the pricing area that folks are working on are, you know, not really possible without these types of standards in place. So let's briefly talk about some other legal tech trends. I'm not going to talk too much about this. First of all, it's incredibly broad. We have a very short period of time and and and as is organized for all of you the links to the standards which anyone can delve into in more detail. Oasis Open is, you know, a place where they're lurking on contract standards, everything from maintaining contracts, you know, updating them transmitting them and so forth and so on. You know, that relies on some XML standards and that transitions us to the next phase of modern technologies. Some of this is done in XML. Many of us consider that to be, you know, yesterday's technology with Jason being today and tomorrow's again that all depends on on on where you sit. You know, I as a practitioner still have certain interfaces that might be an even older technologies. So, you know, it's but clearly the way the world is rest API is, you know, we're writing some systems that we provide to our clients. You know, in things like no JS and jQuery and JSON and so forth. And so that's that's the way the world moving things to a client side sort of world. Notable context on Oasis which is not well known outside of small standard setting circles but they also were the lead in the late 90s and remember working on this a lot and for court filings. So anytime you can submit, you know, like a matter before a court in the US there was a every every jurisdiction basically had its own way to do it is very idiosyncratic or they just use FTP. Now there's there's a wide adoption of these court filing standards so we can capture the metadata and, you know, vendors can build systems that work across a bunch of courts. They also have the coma legislative markup for legislation and they were, you know, kind of part way through a better notarization standard. So that that had finished in time for the pandemic but they've got a nice little cluster of legal standards most of its business standards and Oasis but I think it's notable. If you poke around there that there's a small but thriving legal standards community. Absolutely and I'll close briefly on this but you know, SEC and Edgar have made similar headway with financially filings 10 cues and so forth. And so, more and more, the standards are entering our industry, and more and more people see the common value of them, you know, so that, you know, we can build things once and leverage it many times which is a far more efficient way. Okay, so I think I kind of went on and on about Sally when I spoke about them but just to add a few extra words, you know, it's a standard setting project. It's a broad consortium of members from a very large cross section of industry. And just one quick example as I, you know, as I did research for this talk. I spoke to some folks from that documents and they talk about how they apply the Sally standard to documents in their system so they have a great and better understanding of the type of work and expertise that they have. This has a lot of value to them as they look to onboard new customers, talk about where their strengths and weaknesses might be and so forth. Another example in industry is like we briefly talked about, you know, looking at things like tasks and applying standards to them so we're not talking about apples and apples, you know. And I say I have to, you know, give this talk today, there should be some sort of coder standard that helps people understand I'll be spending an hour or so doing it, you know, not, not a week, you know, but you know that's all those things are important and then they're beginning to expand into other working groups, other countries, there's a Canadian working group so a lot of great work going on there and I highly recommend that you check them out and see what they're doing because I think it'll become more and more important in industry. And finally, I'm going to close with a very brief talk about how standards can be beneficial in a lot of what I do. We have a software company called Zerdict, we have a product called Case Ensemble, it's a legal matter management product, we're not big, we supply it primarily to our legal clients and to a few others. But as we work, leveraging standards is really, really helpful. There's so much open source material out there in JSON, XML and different technologies. A lot of it's stored in GitHub, if I remember correctly, your course materials are stored in GitHub. So you should all be familiar with that but it's an awesome place for source code management, checkout and so forth. So, you know, another good thing to learn about. But again, you know, just a closing comment is, whether it's business 30 years ago with, you know, order the cash and things like that, or the types of things we're studying in computational law and legal, or if it's an application development or software development. So standards are very important thing to be mindful of they provide great value they help you work more intelligently more productively and then higher quality. And, you know, there's always, always keep those in mind to try to be sure that your work is consistent with other practitioners in your industry. So that does or I think I've made it through my slides, at least some are successfully. I know we have an appendix with some links, and I'm happy to engage in further dialogue via questions or any other way. Thanks, thanks so much Ken that that was perfect. I appreciate all the diligence you, you exhibited and putting the slides together and actually conducting some interviews to get up to speed and Sally and other questions. So, you know, for all the guest speakers that are watching this and who are going to present later, look to Ken as your role model for how to do this. And I think just by way of maybe seating a little bit of the of your of the thinking for the students were posing questions in the big scheme of things, you know, law and standards, you know, they're largely not related. Those standards are technical engineering, you know, science, but you know, number one law applies to standards. Very much so a lot of the standards makings, people wrangle over the copyright and patent ability liability and there's a whole industry that you can find out more about a consortium.org and Andy up to grow site that I linked to in the chat. But more even more on point for computational law, there's law that applies standards. So, almost all the law related to weights and measures. You know, like how do you know how much is a gallon at a, when you're filling up your tank for fuel, that's heavily regulated and they incorporate by reference, you know, like test ban treaty nuclear test ban treaty incorporates by reference, very particular seismic regulations, you know, building codes are incorporated by reference we mentioned xbrl extensible business, reporting language is a dialect of XML that standardized, it's incorporated by reference in the SEC's regulations. Also, there's legal technology standards which is really I think the focus of of Ken's talk with Sally billing codes and the contracts. And the law itself, which we foreshadowed a little bit that's when the standards are not our for law so us markup language is what the US Congress uses to mark up all legislation. We talked about a coma is an international standard for that. And the court filing standards and so this sort of crops up there's many dimensions of relevance for standards and standards are so very important, just not only to benefit from the thinking of people about how to engineer a system and how to make sure it's fit for purpose, but also as Ken was kind of foreshadowing a couple of times in the talk for interoperability. So it's one thing if you're building a one off system that you're the only person or a small team that will use. You can be a lot more idiosyncratic there I think with your code, but if you wanted to play something broadly if you want it to be adopted across the market or an industry, and most particularly if you want it to be able to interconnect with other software. The standards can be absolutely essential. And, and we're still lagging behind in the law with standards adoption. And that's a part of I think a reflection of the law lagging behind other industries with the uptake and absorption of technology so there's a couple of provocative maybe partly framing things to maybe catalyze some questions so do we. Do we have any questions that anyone's noted that you would that we should pose to Ken. Yeah, we can go off and order that works. Okay, what do we have to me. So we have from three in a nutshell anyone explain why Jason is better than XML and other formats, for example. Okay, I'll start on that one but I'd love to hear from others as well especially people have really worked with the stuff. So first of all, the word better is super loaded. And so you know it's really like what is fit to purpose. I generally find Jason way better for almost everything I ever do with technology and XML almost never comes up. If you're doing a markdown of a document with the document structure, XML is really good for that. The, the, as XML is itself a sub part of a broader standard called SG ML, which has a lot richer functionality and you can go much deeper. So that's great if you're marking up documents, or if you're dealing with a legacy system that does things with XML terrific. There's also some XML and other parts of technology sometimes it's used for meta, meta information, you know, encapsulating other data. The thing that's great about Jason, I think is that number one, it lets us become more data centered and less document centered. And that's very, very important when you're having dynamically generated information or when you want to break things down to smaller modular pieces, when you want to do analytics, that kind of stuff. It's, it's a struggle to turn something from XML to you know traverse the schema as it were, and to turn that into something that's either tabularized or object oriented or that you could is suitable to put into a database or something where you can start to do analytics or anything with it. It's like really hard. Jason, it's like, hot butter, it's super easy. The other thing is that it's, if you're using a rest API application programming interface and that is the, that's the lingua franca of the web. They all are native Jason. So like that's how the data is structured and serial and serialized. That's a good time for me to jump in because I was going to be my point was I'm going to just keep very general but Jason is part of rest API and part of, you know, 2.0 technologies more even more broadling. And, you know, one of the huge elements of value we see in moving from server side to client side technology as we redo our systems is just reducing calls back to databases, processing more data in browsers, and, and being far more efficient in how we render information to, to our customers to our clients. You know, a lot of legal has tremendous amounts of data, maybe not data per se but large document sets associated with, with cases and so forth and so the more time you can process without going back and forth to the better response time, the better service, the better, you know, acceptance level you'll have with your users in terms of how things work and so forth. So Jason is kind of just packaged with a lot of other things, you know, within rest API and it's kind of, it is just the way to go and I know that's not a technical statement. It's the way to go as you're rebuilding building systems in the future. And just a quick program note. We have a heavy weight with technology and standards in with us today. Brian you listening who really did great work with Thompson Reuters labs in Boston and so much other great work is saying to everybody almost like a safety note. SGM L is way harder to deal with an XML. Don't go back. Yeah, I agree. Don't go back to SGM L just notable that it's there. And it does more. So, do we have time for one more question can I know you got to run soon. No, no, I have time. Sure, go ahead. Okay, so malaria asked talking with lawyers in your job. Did you consider if there are any areas of law that are easier or harder to adapt and transpose in these standards. If yes, which ones are they. Okay, so. Certainly, I mean, depending on who you're speaking with and what type of of practice you're working with. Things can be easier or harder. As an example, lawyers who are working on contracts and agreements for the most part are quite used to standards many of the ones that we've talked about in this presentation. So lawyers who work more on M&A and financial work and with other business professionals who may be more technology astute than others. And or or with those who work with the largest companies in the world are probably, again, more apt to adopt standards and so forth. You know, conversely, you know, maybe lawyers who I don't know work on research or trial matters or, you know, other types of works that are more narrative and, and, you know, qualitative rather than quantitative might be less likely to even understand, you know, the standards is, there's probably lots of folks in the world have problems digesting standards like this. It's not something you can answer very easily but but sure I mean different classes of folks are more or less adaptive to all technologies standards included. So, Ken, I want to thank you so much for the for the talk. If people have more questions, you know, pop them through our questions form which we'll send a link to. And, and, and you're welcome in this class, any time can you ever want to pop in. Just thank you again so much for your, your time and your diligence and for the excellent content that you shared. Well, it was a pleasure to be here and like I said it was a great learning experience for me and I apologize for having to leave early, but I certainly will be in the future and wish you and everybody all the best have a great day everyone. Thanks so much. And so I am going to take the liberty of handing it to Brian Wilson, who's going to set up a really fascinating part of this class that I'm so excited about. So put your seatbelts on everybody and prepare to be, I'm just going to say blown away by some talks that I just thought were seminal for computational law. I know I'm ruining the expectations game here you're supposed to set them low. I'm just setting them accurate. So Brian, why don't you, why don't you set it up for us and, and then afterwards what we'll seek, like maybe one question for each, for each of these talks, hit it. Sure thing so coming up next we're going to do something that we haven't really tried before in this course and that is this idea of lightning talks so we're going to hear from the three TAs Andrew, Megan and TMA on some topics that they've been doing research into. They'll give about a three to five minute presentation will take one question so please kind of be adding those into the chat chat on zoom and then we'll kind of go into a bit of wrap up. So first off, we have the legislative recipe syntax for machine readable legislation. Take it away Megan. Right. Thank you very much Brian. The prospect of machine readable legislation is both terrifying and thrilling. This renewed popularity is owed to the rules as code initiative. The fur for around rules as code was accelerated by the recent OECD observatory of public sector innovation report titled cracking the code. This articulates how machine consumable defined as machines understanding and actioning rules consistently reduces the need for individual interpretation and translation, and helps ensure the implementation better matches the original intent. This methodology enables the government to produce logic expresses a conceptual model. In effect, a blueprint of the legislation. So what is the attraction on what are its limits. I frequently turn to this example. Alan lamented about ambiguity and legal drafting owe to syntactic uncertainties. In a fascinating study he deconstructs an American patent statute, and notices immediately the complexity with the word unless he asked whether the inclusion of unless asserts a unidirectional or a bidirectional condition. That is, does the clause mean a, if not X then why, or B, if not X then why, and if X then not why. Next to Alan exposes an ambiguity that muddies the legal force of the statute. An interpretation of unless as a bidirectional condition raised the question of what not why would mean. In this particular case, this could affect whether exceptions are possible and determining patent eligibility. In short for Alan, legislative language must have a clear structure. If not new, the ancestry dates back to 12th century logicians reflecting on the use of mathematically precise forms of writing. In the mid 1930s German philosopher Rudolph Carnap reflected in a logical syntax of language. His argument is that logic may be revealed through the syntactic structure of sentences. He suggests that the imperfections of natural language point instead to an artificially constructed symbolic language to enable increase precision. Simply put, it is treating language as a calculus. More recently Stephen Wolfram made a similar argument. Simplification he states could occur through the formulation of a symbolic discourse language. That is, if the poetry of natural language could be crushed out. One could arrive at legal language that is entirely precise. Machine readability appears then to bridge the desire for precision with the inherent logic and rule of specific aspects of the law. In other words, a potential recipe to resolve the complexity of legal ease. However, if a new symbolic language like code effectively enforces a controlled grammar. What are its implications as it moves across the legal ecosystem. In particular, its interactions between various legal texts. The machine readable legislation may therefore be regarded as a product that evolved out of the relationship between syntax structure and interpretation. But at the core it boils down to one question. What should be the role of machine readable legislation. Is it simply a coded version of legislation, one possible interpretation, or is it a parallel draft of the legislation, one that has legal authority. Is it a domain model of regulation from which third parties derive versions, such as an open source code to say these three scenarios and of course not exclusive to the three have their own sets of implications. And only in answering this question, what a fruitful assessment of how logic syntax and symbolic language found in machine readable legislation are capable of representing legal knowledge. Thank you. Very good. Does anybody have any questions that we might go to. And if there aren't any I might speed one. So, Leonardo has a question. So this is considered a technique of interpreting text legal text for our matters, while at the same time creating it. And this contact concept is very important in many jurisdictions, including Brazil US and others, and which there's so called this notion of judge made law. How do you reconcile the kind of like ambiguity of judge made law with machine readable legislation. Really fascinating question one of the sort of things that I point to when I say, what is sort of what would happen if you have machine readable legislation sort of interact with other legal texts. For example, if it's related to how it affects sort of contracts and in a way how it affects judicial opinions it's much to do with the conversation around it. And so I point back your question to which version of machine readable legislation are you considering here. I said that it's just one variation so when I talk about the one possible interpretation, because if it's just one possible interpretation, then it means that you can still often refer to a natural language variation of it. And that's where that dialogue between sort of judge made law or you know judicial opinions can interact with the legislation, very similar to what's happening now and there's no change. But if it if the machine readable legislation does have legal authority, then the question is different right. What would the court regard as the one that should take precedent should it be the coded version or should it be the natural language version. And so your question is sort of. What do you think, again, I'm turning it back to you. What do you think the role machine readable legislation is because its implications again are dependent on kind of these various ways to see machine readable legislation. Yeah, there's even an article that I would encourage people to check out called transactional scripts and contract stacks by Dave Muffin and chain and Coney that kind of touches on some of these ideas. So, with that we'll send around some more responses to these questions and then we'll get, and now we'll get to Andrew Dimzalski who's going to talk to you all about the law of automated and autonomous legal entities. Thanks so much, Brian. My lighting round discussion focuses on the laws of automated and autonomous legal entities, some in this class, especially those with a blockchain background, maybe familiar with the concept of a decentralized autonomous organization. My discussion will take as a thought exercise a legal entity that goes one step further than a doubt and automated autonomous organization. It goes into a program which conducts legally significant affairs on its own. The overarching question here is, how would the law treat such an organization. What would the standard be for automated entities. I think that question can be split into two parts. First, how would the law treat liability stemming from an automated entity. And second, would that entity have legal personhood. That's an example of an autonomous car injuring a pedestrian. What entity or entities has tort liability for those injuries, the manufacturer, the driver, the AI. Is there any criminal liability to assign cannon AI have criminal intent. And what would you do with it if it did. How do you place an AI in jail. And finally, the literature is considered many different potential liability frameworks for AI, including strict liability agency law or products liability, or inventing a new framework entirely perhaps a jurisdiction. In his article how to sue a robot, Roger Mikowski coined the term in robot for this purpose, a jurisdiction specific for artificial intelligences. And it relates to the second question on personhood over the years the law has assigned different legal statuses for many non adult human persons. Children, animals wild animals corporate persons even bodies of water such as rivers have acquired legal standing in certain circumstances. What standing should an autonomous organization have should AI have intellectual property rights over its inventions or works. Is there any authority to conduct business as any ordinary LLC, or should we fashion a new legal status specifically for autonomous entities. These are difficult fact specific questions. However, the legal literature has provided several concepts that I have bundled together into what I call a working proposal. First treat autonomous entities as an agent under agency law. The principle, the principle is a human creator controller owner who defines the AI scope of authority into a registry of some kind. If anyone is familiar with the concept of a robot txt file, this could be akin to the AI registry. The AI is permitted to act within its scope of authority but should they ever leave that scope strict liability would apply. But strict liability against whom many scholars commenting on the ramifications of autonomous weapons systems have articulated phrase meaningful human control. That is, those with meaningful human control over the AI could be responsible in a strict liability framework. This would satisfy a social demand for AI responsibility and avoid a potential accountability gap that could occur. Consider the scenario of an autonomous car injuring a pedestrian with no human person being held legally responsible. It is still very early in the evolution of the regulation of autonomous entities but it is a critical discussion to have as it may become very relevant sooner than many think. So, what would be your proposal. How should be a portion liability stemming from the actions of artificial entities. How should the legal standing be of AI, and perhaps most importantly, what implications would our treatment of AI have on the future of law and technology. Thank you. Thanks Andrew. As of the time constraint and because it might be easier to respond to some of these questions a little bit asynchronously I would actually encourage people moving forward to ask their questions in the telegram class chat so that we can be sure that we get through all of the, all of the speaking and all of the instruction for this class. And so with that I'm going to hand it over to TMA who's going to discuss a topic called the algorithmic sensing. Thanks Brian. Hi everyone. Today I'll be talking to you about the future role of algorithmic sentencing and human judges. Recently, as countries like Estonia and Singapore have started experimenting with algorithmic sentencing in their small claims courts, it's become more important than ever before to think about shifts that algorithmic sentencing may impose on existing judges, particularly the trust and standards that apply to the role of the judge. Because the specifics of justice systems around the world very so much for the sake of simplicity will keep this discussion to the US justice system. In the US, it's currently pretty well established that our justice system is riddled with ingrained biases and inequities. People of color are not only overrepresented as defendants in our criminal justice system, but they also receive longer sentences when they're white defendants. And these injustices seem to be the byproduct of human biases and prejudices and sentencing by judges, which algorithms could certainly avoid. The judges today would concede that a mere spreadsheet providing data on past sentencing decisions could help them make more objective decisions. So in this vein, it seems like carefully coded sentencing programs could skirt situations like when an Ohio judge went against the recommendation of both the defense counsel and the state prosecutor to condemn a 55 year old woman who was a first time nonviolent offender to 65 years in prison for petty thefts. It's true that when a man condemned to a life sentence was condemned to life sentence from nearly attempting to steal a set of head shears. It's true that a non trivial objection to algorithmic algorithmic sentencing is that data used to program the code is often incomplete and incorrect, which biases outcomes. But assuming this can be remedied. Society would even be amenable to algorithmic justice. In the United States, judges are expected to be arbiters of justice. The code of conduct for the United States judges states as a very first canon that a judge should uphold the integrity and the independence of the judiciary. Further, they should not only maintain and enforce high standards of conduct, but they should personally observe those standards, so that the integrity and independence of the judiciary may be preserved. The judges be able to personally observe anything and in turn fulfill the standard under the social contract theory, the concept of justice exists because of collectively negotiated human belief. Does the US justice system work in part because Americans believe in the idea of human judges as arbiters of justice, or is it unnecessary. It's clear that the advent of algorithmic sentencing is calling into question the role of human judges as referees of justice. With all of these considerations we went over today, I invite everyone to think about the pros and cons of both human judges and algorithmic sentencing and what you think the best path forward should be. Thank you. Thank you so much to me. I think this is one of those topics that becomes increasingly interesting because of how much it is going to affect everything moving forward. The tools of today, or the tools of the future, or the outcomes of the future look a lot like they're going to depend on the tools of today and so with that I'll hand it over to DASA to preview the next course and then we'll get into some ways to get involved and wrap up. Let me just triple check. Do we not have time for one quick question for Andrew and TMA at this point. I mean I think I can be super concise with my wrap up but it would be good if we could give the meet your shot. I've got a hard stop in five minutes and so I can't go through the ways to get involved if we do the questions and so it's up to you. I have the same hard stop for the next meeting as you do. So I, I'll take responsibility to say what needs to be said, and I want to pose quickly so please be concise. I have a question for Andrew from Dispena and it's, would it not still, so it's interesting point would it not still stumble on the limits of that part of AI that is not explainable. And I think that related to, you know, kind of holding these entities accountable, and you might want to mention the aspects of strict liability in the US context. I do think that it certainly touches upon the strict liability aspect of it but it also perhaps touches upon the products liability aspect of things. And that was one of the, the parts that I considered in adopting the working proposal, but one of the limits that it comes into when you consider applying AI liability in a products liability situation is the the case law products liability requires a demonstration of using a reasonable alternative. And so in, in advanced AI situations you're going to have situations in which perhaps it's inexplicable why AI conducted a particular action at least as far as you know human adjudicating the AI or perhaps it would simply become way to a verbose to go through a situation like that in a litigation context. So explainability is certainly very difficult, especially for a sufficiently advanced AI. And that's why strict liability kind of as a necessary container, I think works perhaps better as a way of having the legal certainty that will assign liability to a human person with meaningful human control over the AI regardless of the factor outstanding thank you so much Andrew, and we have a quick one for TMA an excellent one from Walter Stover does the attempt to make a non biased algorithmic sentencing system by training it on a database that itself may reflect bias against certain individuals in other words by training it on pre on all the prior sentencing. You know, it's, you know, isn't that bad, basically. And so like, what, what do you think about that and can you think of any alternatives. 100% that is a concern with training algorithmic sentencing. But I think that so there's an article in the Atlantic that talked about how judges, when they were able to look at spreadsheets that showed sense prior sentencing they were on and hold able to have more more objective decision making. And so I think that when we're training and an algorithmic sentencing a sentencing algorithm, if we're able to have humans do reflect on you know how objective each decision is and have those decisions reviewed and that data are pruned. I think over time we can reduce a lot of the biases that we see inherent in a lot of the data that's are being used to code these programs. What do you think as a this is on you. So super quick. Thank you, TMA. And so we're going to send you an email today with everything you need to know about what happened and what's going to happen. But I want to highlight, go ahead and join the telegram channel if you can we'll send a link to it. The office hours are turning out to be great if you're free on Tuesday at 1pm Eastern and contact Brian, I mean one of the biggest wells of potential value from this class is doing experiential learning. Brian is selflessly volunteered to actually mentor and organize projects, take them up on that if you haven't signed up yet, you can do a paper, a prototype, right other stuff. And so I'll send that link again to it's not it right it's not too late or is it too late now Brian. No, it's not too late. I would say, you know, fill out the form that you're interested in submitting a project. We've already had a few submissions I'm planning to get through those submissions this this week and get feedback to everybody. The only kind of hard deadline, I would say is if you want to submit a project, submit it before the next class period, and then during the period between class three and class four. We are going to request that students submit a final project and final projects for the scope of this course is different than final projects for maybe where you want to wind up. So last year we had some final projects that took the form of, you know, Google slide presentation, or, you know, just a pitch, and wound up becoming fully fledged papers that we published in the computational law report later on. So I would encourage everybody to kind of find, you know, a way to get something down that sort of represents the kernel of the idea that you're interested in. And then from there, we can go into and work together to find a way to get that to whatever the final place for it should be. Indeed yes so in a sense it's the final start of a project if you want to continue it. TMA, could you bring us home. So guys, this wraps up our second class for this year's computational law report. And, like last time, if you have any questions, please email us fill out the Google form or message us in telegram. Okay, we'll see you guys next week. Thank you. Bye take care.