 Yeah Hey everybody, it's Brian Wilson here for the computational jurisdiction workshop session and The 2019 MIT computational law course. I'm joined by Brian you listening Thompson Reuters labs and our dear colleague does a greenwood and we're going to be kind of Meandering around this idea of what a computational jurisdiction might look like as Per a video that Kristoff Perera from GE posed at the 2017 MIT legal forum And so we're going to talk about some of the ways that that could be designed You know what it might look like as far as how it is organized And then different mechanisms that could be used in order to ensure a certain level of Integrity just with regard to being able to control the things that would need to be controlled and ensure the outcomes that we would Find desirable. So as always, you know provide your feedback to this session in our pigeonhole widget and We'll wait a couple more minutes, but Before getting going in earnest, but in the in the meantime, you know Brian tell us a little bit about yourself. Yeah, sure. So Brian you listening I run Thompson Reuters labs Americas so I have a group of data scientists and data viz folks in Boston we deal with Thompson Reuters that provides Information for professionals particularly legal professionals through the Westlaw Platform and other legal products as well as tax and of course, we've got the Reuters news agency and then I've also Got a small lab in Dallas that's data scientists and blockchain type folks So we're interested in in all kinds of distributed ledger applications for legal and tax and Yeah, happy to be here as always Yeah, yeah Yeah Hi, and just by way just a little bit of context. This is actually this is the session of Brian and Brian but um, I just want to vector forward that the underlying idea was that of Kristoff Perera as you said from GE and That was that and in the The idea was such a good one that it's engendered a lot of conversation a lot of Brainstorming in the year plus year almost year and a half since they had and that's probably why we wanted to continue to explore it as a breakout session now and Brian had the very wise idea that maybe I should tell Kristoff We're doing this and so I did do that and we had a nice chat with them last night and just to bring everyone up to date Um, Brian Wilson and I Kind of briefed him on what the plan was he was delighted with it, which is good and we asked him As per your suggestion, Brian, you're listening you have yet any further thinking or any any suggestions on Questions that maybe we should focus on the main thing that he says well, it's it's It's good to once you kind of discuss the idea and kind of get your head around it To focus on what is the smallest Appropriate one could deploy and test the idea and Brian had some ideas Or you guys had discussed like a cruise ship is an example and burning man in some ways is an example and and he blew my mind Which probably went one to pop in this session a little bit of maybe one way you could test it would be with a burning man camp That did this and as soon as he said that everybody was thinking well So there may be an opportunity for a fun That's one fun at least example of a way one could take some aspect or aspects of the idea and begin to build out like a contained Environment in in time like a week and in space and think about and What elements would be engineerable and what kind of data could you get and how could one go forward? That was sort of a somewhat whimsical fun Idea, maybe you'll come up with others during the course of the session. Yeah That's great and said we got some good feedback So I think to start we put together Notes and we can kind of walk through those those notes just a little bit to kind of provide context But then really from there. We'll start diving into questions so for me and on in my experience as a law student I got to work under Kansas City, Missouri's chief innovation officer on the smart and connected cities project and That project was a joint public-private partnership between Kansas City Cisco and Sprint and the goal of it was to embed a really large sensor network around Kansas City Streetcar corridor and use those sensors in order to Kind of gain insights about you know the air quality transportation flow different aspects like that, but then also use it to help do things like turn the Lights on when there are more people around late at night so that you know You can help hopefully prevent crime or just make the city more comfortable and user-friendly And in that project one of the one of the best conversations I had was with this retired This retired nuclear systems engineer who was asking You know trying to hypothesize different systems that were already out there that were very measurable in the same way that smart cities would be measurable smooth and It to that end we talked about you know, you can measure everything on a cruise ship because you know everything that goes on it and With that it's like okay. Well, this is this very large and kind of portable city It's in a way a portable computational jurisdiction. I know a lot of cruise lines now They have they have the same sorts of issues with information security that a computational jurisdiction would surely Have at least there would be some overlap there But I think you know having that sense of you know, there are going to be some information security requirements It definitely is a good backdrop against which to set the the computational jurisdiction idea and then also I think Some of the things that we touched on yesterday in the course as far as having you know some Security principles by design that are kind of built into the framework from which you're starting this kind of gets to this larger idea Okay, if we're having a computational jurisdiction We need to start from a series of principles and then use that series of principles to Design the technology and the legal applications around it so that they can function in the way that we want so what in your mind would be some good principles that We could have to start with I Wonder that's open to Anyone can we see the pigeonhole? Oh, yeah so Participatory decision-making seems like a really interesting one because you know I know with especially quadratic voting now that you can do with some of the blockchain apps and smart contracts that allows for some Some really creative possibilities Especially With figuring out which permutations of voting schemes you might want So that could be that could be a fun one Yeah, yeah voting scheme seem like a Really natural way that wouldn't require too much in terms of sensors. Yeah It's it's pretty straightforward and and I think that also if we take one step from voting schemes you could get to the idea of self-sovereign identity again and Just have your self-source data producer interacting with all the other computational apps that are in this computational hypothetical jurisdiction And going back to the cruise ship Analogy, I know I think it was carnival announced their their passport thing It's like a wearable device that you can use to pay on the cruise ship in whatever currency that you have You can verify that you're over 21 you can verify that or whatever the drinking edge is wherever you are so 18 In a lot of places you can verify that you've signed up for certain sessions with whether it's a spa or you know Special movie, I don't know Haven't been on very many cruise ships recently but I what what are some other applications that you could think of that might be be computed out of a wearable just like Taking away the Self-sourcing data part of it for a moment, but like just like what are some cool things that a wearable could do in a computational jurisdiction Yeah, yeah So one thing that that that calls to mind Is a very rich thread of the discussion That followed Kristoff's presentation and his kind of you know, good-natured challenge to Co-think that how to build this prototype idea and and it was what kind of data What kind of data would be appropriate and much of it most of the discussion was around Personal data that would come from a phone or that people would self-enter. We talked about game of creation. We talked about And then the you know, what is the kind of value proposition on so what sort of data would would be relevant to and help to Enable a computational legal system like parking and traffic data location Tax data You know transactions Etc. Etc. So yeah, the data that's coming from individuals was very much part of of the brainstorming last year Yeah, I mean parking seems like a Straightforward because you can actually do all kinds of AB testing with these smart parking meters, right? You could have a Through your phone you actually get They tell you what the policies are right? Oh, I don't need to pay on Sundays. Oh, you know And then you could so you could definitely do lots of maybe maybe they are maybe the city of Boston is already doing that These places and you could then see, you know like well, so how How long do people stay in parking spots? How full are they? How easy it is to find, you know in certain parts of the city Yeah, maybe that would be a actually we can proactively go talk to some of these parking folks and Talk about parking policies. Yeah Well, and depending on how big that network is to you could monitor things like commute times to see how far people are commuting from You know what? for different areas of the city whether it's socio-canonic or otherwise you could you could start getting an understanding of You know what the factors are that are contributing to maybe undesirable outcomes and start Proactively trying to address those as well Yeah, you can play with you know commuter taxes and like I'd be willing to pay a tax to You know In to drive into the city if I was more or less guaranteed that there would be yeah Yeah, you can go ahead and reserve it before you even set out. Yeah at a certain time and then have You know, I know I Think it was open while they did it. They did a really cool demo one time I think it was last year when I was in New York for the competition along watching festival there, but they were They had a live example of a sensor that they could embed in a crate for shipping and with that If you had some sort of perishable good like flour You could monitor the humidity of the crate and if it passed a certain threshold There would be certain penalties imposed automatically for however long it was at that Temperature or that humidity level and then you could use that to kind of have these contracts that are automatically Executing based on that and I'm thinking, you know, if you spend a certain amount of time, you know in that parking space It can vary nominally just Give you that liability for you know Whatever amount of time that you actually had instead of you know the the system that we have now where you put in So many quarters for an hour or so many quarters for two hours And then you maybe wind up spending way longer there and you get an additional penalty Or maybe you suddenly get called away and you've wasted that, you know 80% of the money that you spent So I think that ability to kind of reach a deeper level of kind of like customization appeals to a lot of people because it's actually Making it so that the Transactional experience is one that actually matches what you're like the the the cost actually matches the services of benefit that you're getting And so what so so looking at some of the ways that we might test this I think In reality one of the easier things to do could be to Just have something like a burning man camp where maybe you monitor a handful of those outcomes Maybe maybe you start with five maybe you start with kind of like the mobility in and out of the camp Then you can look at and you can track that based on sensors You can track, you know How longer people in the camp if people are wearing the sensors out of the camp and going back and forth to other places Can we come up with a way to monitor that? From there you could also monitor. Maybe the the goods that are going in and out of the camp if there's a system for that What are in your mind Brian are there any Issues that you could think of just like with a can't like a regular old campsite that you could start monitoring that could Kind of in a way be analogous to something that you could use in a larger scale jurisdiction Yes, so Just to put the campsite up on hold for a second. So I had put on telegram This this paper that we had done in my previous job and which might actually lend itself to these kinds of experiments and it was about Setting policies so, you know, you're on you have a chat Application and you want to set policies about who you will exchange information with and you can also set policies about Access to data that you have so access to Databases so there's you know, like certain people can get certain kinds of information about personal Yeah, sensitive personnel sensitive information and not and that was kind of interesting because you know is this decentralized thing where everyone could have their People could have different policies and then you could see the behavior that emerged of who could talk to who Yeah, and who I'm you know communicated with with who and so on and Definitely work through issues about You know information exchange among bureaucracies and things like that that way Yeah, I I think that gets to a larger issue that we've talked about in some other sessions too of going from something that's You know Law-centered to something that's human-centered so we're taking laws and regulations which by their very nature you know at the at the beginning we're centered on the law and It's moving it to where it's something centered around human interaction and I think that as a principle is Is something that we all here can be very big fans of because now we're living in an environment where we're producing Data we've got data exhaust that covers all of these different things that have All these different behaviors that we have on a daily basis a lot of it that we don't even aren't even cognizant cognizant of and that that data can be used to you know reduce a lot of the burdens that we go through in the legal process I You know the legal processes that we have now what lots of them were designed for you know not having Very much trust within the system so Acquiring lots of papers to go back and forth to lots of different people so now that we can kind of Automate a little bit that that degree of trust whether it's through something like multi-factor authentication or something else That serves as a really exciting possibility. I think for you know just the way that The the concept of like a a justice layer or a legal layer to our society might Might be imagined So looking at some of the questions now What could be some of the best principles to create a computational jurisdiction where self-serving interests do not Compromise coherent policy-making and rules and then how do how are those included? For the interests of everyone US legal system so speaking I can I can quickly reference here the I Linked to it in our notes, but the I Think we can start with the ten principles that they've got for burning man Just as a jumping off point and then move towards you know additional kind of Principles for building something that serves to be an area where people live and communicate and are Interacting with one another. I'm so I will throw these into the comment of your question Just so that everybody can see them Is there an And now that the terms of the origins of the United States that's a really cool principle to have welcome and respect Strangers know for participation, so that's I think a good principle that we could Radical inclusion is the second one or that was the first one so gifting is the second one so they They operate in this kind of gifted economy where you know the value of the gift is unconditional and When you gift something to somebody else, you don't really contemplate some return or some exchange And I think that would be a little bit difficult to have in a real place just because of the nature of the economies that That exist That would be certainly something that would be fun to test out and in a lot of ways I think we are already testing some of those out in a computational sense like we have records of transactions and we have Centralized institutions and banks and stuff that you know kind of do a lot of this computational work for us So maybe you know that exists as sort of a layer on top of all of us I mean these kind of remind me of if you go back to Rawls and you know the the kinds of principles he You can sort of see Rawls The whole idea of the you know the veil of ignorance is being sort of a computational Law set up right so we're supposed to imagine the rules of society where people don't know what what What role they're going to play and so they adopt these these rules by you know Trying to think through well if I didn't know what role I was going to play What rules would I adopt and and he thinks that everyone would choose rules that fit certain criteria You know that the you would opt for the rules that have the greatest benefit the least advantaged and yeah Those are actually like very straightforwardly sort of computational kinds of processes so you know so you can Even so that Rawls's critics have different sort of views about well It's not just the greatest benefits the least advantage, but you know like Maximize ability and things like that. Yeah. Well, I think to the the kind of ideas Around utilitarianism just as a philosophical concept You know, we're starting to get to points where you can actually measure what is the most utilitarian thing The most utilitarian outcome for a given situation So I think in terms of like figuring out the the economy layer, that's definitely something that could happen Looking at the third third on the list decommodification This reflects the idea to create social environments that are Unmediated by commercial sponsorships change actions or advertising so basically focus on The human-centered kind of interaction of things and and I think as a guiding principle that is a positive thing Obviously there are you know greed can lead to bad things But I think that kind of is another balance that would have to be Sort of figured out in a computational jurisdiction and maybe you could kind of code that into system rules and some of the ways the smart contracts have emerged that Or or even cryptocurrencies like there there was a cryptocurrency I forget what it was for but it was for some something that was like part of the Proceeds of all the transactions would go to like sustainable earth and in in the Book by Douglas rush cough throwing rocks at the Google bus. He talks about you know the different kind of outcomes that you could peg a currency to and So in his mind or in the book he talks about how you could tie a currency to the production of the timber production of a region or the some Natural resource production or gold or all of these things that over time people have kind of experimented around with And I think there might be a way to like figure out if you if you were to design a computational jurisdiction Yeah, you could design it so that it included a certain amount of decommodification in it in kind of the same way that Zoning prohibits certain buildings from being added to different parts of cities You you would kind of have to have rules like that. I think But I that does point to an exciting possibility of a computational jurisdiction in that you could actually design that into the rules of the jurisdiction So that maybe you know, it's not a very big jurisdiction But every time you go to the your small jurisdiction you're having This decommodified experience where it's more focused on the the human interaction that's there and Moving on down there's and we can kind of Skip through some of these a little bit in an effort to get to more of the of the questions but there's a lot of focus around community a lot of focus around responsibility and a lot of it seems like a lot of focus on sustainability and as far as You know guiding principles. I think you know having people who are committed to you know, putting forth an effort and Making sure that people are responsible for Different levels of participation. Those are those are definitely good things The the the civic responsibility aspect of it is I think something that people always talk about because You know Everybody gets mad at people who don't vote, but then it turns out such a majority of Are such a large group of the population didn't doesn't vote for things And it's like well, how could we how could we design something that holds the people who want to be here accountable? You know and maybe they're creative ways that you can do that I'm sure By having something that would be computational in nature, it would be a little bit easier you could improve that user experience so that You know it it's connecting People to like the outcomes that they're looking for a little bit easier Yeah, so I think the Kristos idea was You know, there's There are you know sort of political science Experiments right that you can do you can say okay? Well You know we found it if we move voting to Sundays that it's more likely that yeah vote and so on but You know I think in his original discussion the idea was that well, you know you can try out a bunch of different policies and this kind of Synthetic sandbox and then that led people and immediately in the initial discussion to talk about well You know, you've only really got data about what people have done under the actual policies So you don't you know You have to make a lot of assumptions about what people would do under other policies and or the idea was that you Have to have like, you know simulations and agent based in simulations and so on which gets into you know pretty complex stuff So it's one thing to measure just kind of model like how people walk around in You know public areas or something like that. That's kind of straightforward. Yeah, like more at a very Complex behavioral Layer where an agent is thinking about Well, I would act this way if I thought that you were gonna do this and yeah, I like it's very complicated sides of things. Yeah, I It's funny as you were saying that I thought, you know, it would be interesting What if we had the United computational states where there were all these laboratories for Figuring out all the different sets of computations that you could have and you could have different people working in connection with other people in order to achieve kind of maybe a unified set of outcomes for That that brought them together but in different and measurable ways that could be used to you know If we if we decide something works great We can just fork we can just let other people fork the code of our computational state and it can be used to You know improve outcomes everywhere else That would be pretty wild thing I think looking at looking at some more of the questions I Think we're starting to tap into some of the more Kind of like the guts of things like how how do you start Mac mapping out like legislation to behavior and how do you start? like drafting laws that maybe are built on you know the Computed data that comes from sorts of behavior or kind of doing that inversely through Like setting out a mission of what desired behaviors are and then trying to come up with different strategies for achieving those Trying to think of the best way that you could do that, but it would probably have to be under the context of you know, we've got we've got a group of people here who are trying different stuff and you know, maybe I know with some of the With a lot of the sharing economy discussions that happened that have happened in the last five years Different people brought up the idea of Kind of like sunset regulations and regulations that are more temporary in their nature Or that require Like revisiting those ends after a certain period of time. I know a lot of the every a lot of the Everything that came after 9-eleven all those years ago. That was that had like sunset clauses and that required Like feedback to it and making sure that it is actually setting out to do the things that It want. Yeah, there's the Patriot Act. I just couldn't think of it but you know, I think you can you could start to do that within like a single jurisdiction that way a little bit so that you're You're actually getting data about, you know, what the problem is and then you can use that data to kind of periodically Like see if to what extent are we reaching our mission and if we aren't, you know, what are key factors or any of the Performance indicators that you know show that we aren't where we're supposed to be yeah, I mean I I'm not No, 100% sure about this, but it seems like You know recent legislation does sometimes include, you know requirements about the data that's yeah That's gathered that can be used to evaluate it, which you know is important because then obviously if you don't bother to yeah Get the kind of the data you can't do an adequate evaluation, but Yeah, so that would definitely be an important principle I think to yeah See More of these are whose it's got comments. Yeah, so so in one of the comments Nicholas is talking about participatory technologies that can assure democratic validity And I think that's kind of been muddled a little bit just This is a quick reflex reaction, but I think you know The concept of democratic validity, especially in the context of you know All of the ability to distort information and have information Falsely be sent to people who don't know that it's false or like ensuring that there is a level of data integrity for The people in a computational jurisdiction, I think that would be something that you would have to You would have to definitely address and I think that's something that people still don't know how to address I know I know that you know that kind of validity you could even take it a step in a different direction where Looking at you know, how can we how can we protect people from like hate speech? It's like there's not an easy There's not an easy fix to this This is what like some of the best and brightest minds are like trying to figure out But it's not like I just flip the switch for you know, no hate speech on my website and Boom, it's gone you've got Many factors to deal with and I'm sure the factors change as the social context of you know Everything that's happening in the society changes too. Yeah, so I think you'd need to you know in the in the Computational jurisdiction you'd need to control for kind of outside messaging and in the ability to Control for that. So you'd want to take topics that were sort of neutral as far as the ability to you know for And Determined cohort on Twitter to kind of change people's minds. Yeah, well and too I think one thing that would be Quite helpful in a computational jurisdiction is that you could have the social network of the computational jurisdiction and it wouldn't be tied to the It wouldn't be as heavily tied to this Huge network of people it could just be for basic levels of communication between people in the community And I think having social networks that kind of fit you know a specific purpose or are very important and Not an endorsement, but that's one of the reasons why I think telegram such a cool idea is just a channel for Communication is because you can invite whoever you want to it It's very easy to set up and you can kind of have discussions whether they're project-based whether they're you know based on something broader like news or Just kind of feedback you can you can set up your own bots that help you monitor different things you can Input commands that say you know what what's happening here and you know that could very well be Something like that could very well be a tool for you know ensuring that you You're maintaining that level of data integrity in this kind of nebulous Lee abstracted idea that we're concocting here Which is a Senate that's a sentence. I never thought I would say in my life, but It's a fun one nonetheless 5-minute countdown. Okay So we have just been told that we've got five minutes. So if anybody oh, we've got two new questions even So we'll kind of walk through those and then wrap up a little bit Oh, this is a this is an area that we haven't touched on very much, but it's talking about in the big data context Figuring out how to take a database for court decisions that periodically reviews Some of the hate speech so maybe you know We talked about kind of breaking it down to a very granular level where you know You kind of insulate yourself a little bit from the sorts of of these sorts of The sorts of bad Outcomes that you don't want to happen But you could also do it from the other way where you go super big at the 30,000-foot level and you start looking at okay in if I connect my Computational jurisdiction to your computational jurisdiction and we aggregate all of our information Let's use that as a way to kind of Mind the language and do some cool machine learning stuff on what language means in different contexts and sentiment analysis and natural language processing in order to come up with a Best metric for you know, this is definitely hate speech that could be hate speech considering on the context and it goes to a Specific party or it's not hate speech and we're good Yeah, actually, so I looked into this a while ago with hate speech per se, but looking for Examples of threatening speech in in like Decisions and so on to do exactly that machine, you know machine learning on it and for one it was really really difficult to find Examples yeah Even a people who have been convicted of threatening people online Which you would think And then but there definitely wasn't enough to really train something that something meaningfully. Yeah, exactly. So and if you have any Any leads on that that wouldn't be like data that could be used on the on the hate speech front or the threatening speech front They're often similar Okay, and then to the last question Talking about the interplay between ethics values markets architecture and legislation so like what can we learn about the this this kind of Interplay between all of these pieces that are necessarily connected and kind of the way that Who was it yesterday talking about the Larry Lessig's Coda's law book where he talks about that convergence of Bill Rosenblatt. Yeah, Bill Rosenblatt's discussion yesterday about you know well all what they were but the the convergence of kind of these four issues are of four issues Many of which are included here but how what does that interplay look like in in this jurisdiction and I think You know, I don't know that I've got a specific answer to that, but I think you could start playing around with ideas of Kind of like commodify at least quantifying them into measurable things so it like looking at that idea that Douglas Rushkoff has where you can sort of tie a currency to specific Commodities you could tie a computational law jurisdiction and I would think to Like as a community you could tie it to you know Certain types of crime maybe like a calculation or a quantification of different crimes that happen, you know How the markets are doing? What public repairs need to be made and then the overall satisfaction of the city from the people who are in the community, so you know may maybe having like a Improved feedback loop for you know getting citizens to engage and say you know what they like what they don't like what they would change if they I think you could kind of architect something that does you know Meet the consensus of everybody who's within the network. Yeah. Yeah, so I'm the David laser over at northeastern has a book about designing like Participatory democracy You know like forums for people to Discuss things so that it doesn't become just like the most Sharp elbowed people showing up at national town halls. Yeah It's it's not the the Loud minority of people who's you know feeding back into this stuff Yeah, and and you could you could even I think tie that like have like a kind of Make that a requirement of being in the computational jurisdiction is that you have to provide a base level of feedback about you know one thing in Such-and-such period and use that as you know how you're going to Make sure that the the city and the that kind of like commons area that everybody is a part of is Actually commons based on the feedback of everybody and not just commons based on loudmouth Brian's idea about what the comments should be Well, I wasn't talking about you Okay, sorry to tell The the the Brian squared lecture But yeah, this is this has been pretty fun And I don't know that we have time to get through more of the questions, but we're we're gonna kind of assemble report out Directly to Kristoff per era for this lecture we talked to him on the phone last night And that was one of the things that he talked about being interested in and if we if we should decide to do the computational law Stational states burning man camp network Perhaps everybody will be invited. We should do it a burning man, but it should only involve parking spaces. Oh Man, they force people to get dressed for work and commute In the desert and park their cars and that's their only activity They would probably I think the emphasis then would be on actually burning those people and Getting back to the fun. But no, this is this has been this has been read everybody who listens and online and contributed Will have I think the feedback which it's gonna live live live with the comments for another while So if you're watching this asynchronously from somewhere else at some point in the future Feel free to provide feedback and hopefully we can get to some more of it But yeah, and if you want to put any kind of useful resources or links or anything in the comments Please do that would be super helpful and we will be monitoring this so thanks again to everybody Thanks to Brian. Thanks you Brian. Yeah, see you guys next time. Thanks Mila