 Great, and we're live from Pivotal Labs here, and about to be introduced, here we go. Okay, cool. Okay, great. Thanks everyone for commenting today to our Tuesday Tech Talk at Pivotal Labs. I'm Penny, and I help coordinate Pivotal Labs, if you're not familiar with this space, it's like I'm offering something for me, and you can have me to talk to you about it. But right now we're going to hear from Dazza Greenwood. Did I say that right? Yes you did. Awesome. So Dazza is the founder of Cinex.com, which is a routine provider of professional control duty services for digital business and trust networks. So Cinex.com enables organizations to deploy cross boundary data flows, transactions, and other services using integrated business, legal and technical architecture and design methodology. So engagement for tailored to-purpose and leverage innovative solutions for digital identity, data policy, and technology strategy. A little background is at MIT since 1997. Dazza held various academic research appointments, including business scholar, lecturer, and scientist, and Dazza developed repeatable methods for integrated business, legal, and technical design and development assistance or services that were moved over decades of private consulting and released under open MIT licenses. Currently at the MIT Media Lab, Dazza's research focus is on integrated legal and technical code enabling trusted identities in the cyberspace, reliable big data analytics, and legitimately fair personal data sharing. So that's a good start for the intro. And so he's going to get started by telling us about his most recent project. Yeah, hack. Hack. Oh, it seems like your mic might not be on just yet. And then after that we'll have our cue in it. Thank you very much. Can everyone hear me okay? Yeah. Great. So this is a project that is open source and I'd say very exploratory from the MIT Media Lab that we're calling IAuth. It's a project that has grown out of the research work of Professor Sandy Pentland who runs the Human Dynamics Lab and basically the focus is on use of big data to gain insight, especially into human behavior and groups and population and trends. So you could think of it sort of like using epidemiological techniques or I was advertising techniques. But let's just say from a approach that he calls reality mining. So we like real time data and we especially like data from individuals. So unlike most social sciences, which are definitely quantitative, you know when I was an undergraduate like I had a clipboard and surveys for Psych 101 and we had an economics that did regression analysis. Certainly quantitative. The difference here is that the types of statistical models and analytics are more like what we would think of as modern data science and less and then specifically based on the models like machine learning or derived from the data directly and not from an assumption like rational actor in economics or psychological model like Freudian or something. So that's why we like the data right from the people. We like individually identifiable data which also is somewhat shunned in social sciences. Typically for research the game is about anonymization and aggregation and perturbing the data so that individuals are not identifiable. So just to give some context for this IOT app, the project that started it was in the Media Lab where I spent part of my time as a researcher not the civics.com consulting that Petty mentioned. In the Media Lab DARPA granted a project to Professor Pentland's group to develop a method for the Pentagon to do a better job in identifying vets returning from the wars who were suffering from depression, post-traumatic stress syndrome were the two key ones. So they threw a lot of money at it in different ways. Most of it more traditional clinical. Some of it went to the Media Lab to do something more exploratory and it was to basically work on a platform out of our lab called FUNF which gets sensors, all the sensor readings from Android phones and puts them in a nice single package that's easy to work with, no wrangling needed. From that it turned out that it was demonstrated that it was possible to have a very high degree of predictive confidence when depression or post-traumatic stress syndrome were present. Partly because these are behavioral. They're based upon observation. There's no blood test and the phones are really, really good at creating and preserving and transmitting data about behavior. So for example if you think, at no point did we get the payload of email for example or text but just the metadata showing patterns of contacts with others is extremely telling about people's mindset and how they're doing and their trajectory in a social environment. GPS is great to see a great deal and accelerometers and gyroscopes can tell you a lot about how people are walking. With early onset of Alzheimer's that's another important area. Slight changes in the way people walk. The way people walk is very signature. It can be a very, very good indicator of early onset of dementia or Alzheimer's. This can be picked up with the accelerometers and gyroscopes and phones very easily. Similarly for the DARPA study when people are not leaving the house so much when they're not communicating as much with other people as you can tell from the metadata and some other tells like that of a fairly thin slice of the data turned out to be very predictive. So the game that I was playing as a former lawyer in this project was basically all about how does one have a valid legitimate method for the Pentagon to get incredibly intimate data about that's after they're back from the war. In some cases they'd be civilians at that point. It almost sounds like the kind of thing that you wouldn't want and like some dystopia where the government has a bunch of information about everybody and they're trying to help us. That was a pickle and it required some really deep thinking and eventually we basically embraced an idea from Sandy Pentland which is what he calls a new deal on data. The vision is that in order to get the most out of the potential of big data we need to use individual identifiable data. That's a big part of it in fact. The only way to do that legitimately is where people own and control their own personal data. A radical that may sound by today's standards where almost all of our data is fractured across some third party systems and very literally when we click the buttons, accepting terms and conditions that we are assigning and transferring ownership or the equivalent of ownership rights and we can't really track where it is, we're not sure who's looking at it, we're not sure what's being done with it from that point forward nor we're necessarily sharing in like a fair value exchange and more of the point, it's kind of quick sand for the companies and government agencies that are relying on it. It's hard to, having been on the other side of that many times, it can be hard to get behind bold heavy duty action premised on personal data that maybe has if illegal or political basis to have because you're never sure where a bad headline or a lawsuit or a political dust-up is going to come with personal data. It can be like a third rail. So what's truly needed to transition to make the most of big data to really come to a digital footing in the economy like big as well as for every company is to be able to have a legit basis. New DO1 data provides that and it's actually not so different from the laws in Europe where people have deep rights to the personal data. We don't have a privacy directive like Europe and so what we did for the DARPA project or what I did was came up with a contract-based framework. So basically every party in the contracts had a legally enforceable by contract even playing field that reflected similar rights to the fair information practices the kind of rights you'd get if you say you had HIPAA covered data or school records. The school can't just give your transcript to anybody or your disciplinary record. It requires consent. Same with your medical records. Same with fair credit reporting at financial privacy and it turns out it's the same with quite a few bodies of records in the United States since Nixon's sort of implosion. There were 74 onward for several years. We had quite a blossom of privacy laws that were actually similar or better in my view than the current European privacy laws giving people very complete rights to their personal data in terms of control, knowing where it is, being able to ask for a copy of it, being able to, if there's something inaccurate, basically correct the record and all these are basically good information hygiene. So proving from some of these fundamental rights, fair information practices that they're called or the wrapper that we use in Sandy's lab is a new deal on data. It's actually a pretty old deal on data. Every party agreed to those terms as though it was a statute, but they did it in contract. So as I was trying to figure out how to really do that, so it wouldn't cost too much and it would reflect and support existing business practices and systems and the same technology that we're using already, it became clear very quickly that the pivotal point, if you will, was the OAuth 2 tokens. So it's that token exchange where all the action is. That's where you can get full alignment of the business legal and technical dimensions of the creation of new value, of the creation of a new asset class of personal data used in the aggregate to do some pretty important things. To save a lot of money, potentially make money to make value. And it also aligns a point when an individual has given a literal, direct, actual consent. If you think about it when you see that little thing pop up or if you're a developer and you've created them, the action required is like a physical action. The user, like Moses or Klicks, there's a user action required to consent to terms that the user sees. That's legit right there. The terms we see aren't always that terrific. Frequently, they almost come across as an afterthought of collecting them for years now and looking at them and they're not always carefully construed or reviewed. But if you look at Facebook, Google for sure, Twitter, all the big companies and many smaller companies like startups from our lab that do clinical use of data using funds, for example, those terms are very heavily negotiated and they exactly reflect what's in the contracts and they absolutely reflect the fundamental business deal, the deal between the parties. So first of all, for that DARPA thing, if you want to look at those contracts and the system rules, those are available for free. They're a bit heavy because the parties were kind of heavy. But the underlying design concept was good and felt it had legs so it kept noodling on it and it was basically to come up with a simple, well, a design pattern, a reusable design pattern that was generic that would permit any company that was, let's call it a resource provider or an authorization provider in any individual that was utilizing OAuth2 to grant access to protected resources, which is what it does, could basically know in advance that that would be the moment that you should be careful about the words to make sure they line up with the deals and to be careful with logging so that everybody has a real log of what happened exactly and to be careful with maintaining the endpoints so that if you, for example, do a token revocation, like we want that endpoint to know when a user is revoked authorization from either side, it's available if you look at Google or GitHub and others, like they will give you endpoints so when, if you have an app and you revoke, and a user revokes authorization for your app with Google or GitHub, there is an endpoint that you could if you cared to let Google or GitHub know that the token's been revoked, that authorization's been revoked and they will cancel it. A lot of times we don't necessarily do that because no one's asking us anything so long as I separate the authorization on my side, we're good, right? Sure, and as long as we just do it any sloppy words we want in that moment and the authorization we're good, right, yeah, I guess but the opportunity is to be a little bit better at that and to actually do two or three things right with logging authorization, logging revocation of authorization and lining up the terms with the contracts and the deal if we just take a little more care and involve business, legal and technical people just a little at the same time, not a lot. It's possible to unleash a major source of untapped value. Here's how, so I wanted to show you now something that we hacked, so this has been more or less an idea for a couple of years with very not great prototypes and proofs of concept that have done grad students or just whoever cared about it, no one's paying for this at this point and so it's been hard to get things together thanks to Pivotal Labs and doing a TED talk on this that can be very clarifying in the mind thinking, well, gosh, what am I going to show? And so we have hacked something together up into three hours this morning and by golly, I think it actually works and to show you what we have done we've got a Node app that demonstrates what we have and so let me just go to Akshith for a moment. Now I should say there's two volunteers here so a lot of the work I do, you would call Civic Hacking so someone from Code for America is here and Massachusetts Legal Hackers are I think following online and I just love hacking as a volunteer, a lot of good things can happen and here's one of them, so Akshith who lives in India was a participant in January short course I did on Blockchain and Legal Intensive and Gabriella who is next to Akshith here can you see that? Gabriella also was a really great contributor and helped me run the class she's here in Cambridge at this moment so here's a couple, these are not MIT students but they're behaving like great MIT students with the code and Akshith do you want me to just hit the video of the Node app that you brought together and then answer questions live? Sure. Okay great so because if you rely on the live demo from India it will never work, we've taken the precaution of making this video, of course it is working now but we've got the video so let's play the video. Why is the Blockchain is quite soft? Can you hear that volume? Yeah. Why is, oh I think I, it's funny it was lying on my computer already so I think what I'll do is I'll just kind of talk through it a little bit so what you, is this paused? So what you see here is a front screen of an app that we put together that lets you log in with GitHub or with Google. Okay. And the next thing we'll implement is log in with MIT which also supports OAuth 2 now in addition to Kerberos OpenID connects specifically. What you have, this one figured, yeah good. So once you log in, yeah here we go, once you log in, you see a little screen that says basically I've taken the permissions that you get and I put them at the top of like the user's profile. So if they want to know what they've agreed to rather than digging for like usually like an apps permissions page it's under a settings page, it's under account management page, it's under preferences page, it's under hard to find frankly for most people many people don't know it exists that you can review the apps that have been authorized from a service. We just kind of put it right up front and center and at near the top of the page and the idea is to see what it would be like to delete the terms and conditions delete the privacy policy, I know it's scary but recognize that they don't work, they're broken. People don't read them if they did, they wouldn't necessarily be better off, they're hard to understand, they're hard to parse I have trouble reading them and replace it with something that could work, something that could work could be something like what you see with a one sentence human readable summary of an authorization that you can click on and understand more about the nature of the resource and you can get paragraphs if you do click on those things what you could do and I've done this with some city clients in particular that are trying to have better plain language and understandable agreements for click through for citizens you can basically just bottom line the two, three, four or five key terms like we're not liable for anything you own the data you agree to share data you can go to online parties and then ever note as people agree to there was permissions, you can have a dynamic contract that actually does demonstrate some of the obligations of the parties at any given time because that's what happens when you look at the currently authorized apps you should see a then current listing of the apps that you've authorized right and when you revoke one it disappears this is basically a screen let me just back up for a moment and say the architecture of this was absolutely not built so that we could just demonstrate something that we could demonstrate with paperware or PowerPoint slides it was built to have three, four fundamental capabilities that should in principle be able to be standardized, formalized built upon and scaled because they have the right things, the right business legal and technical things one of those things is a capability for people to in a sense own their own identity at some level and so when I say own their own identity I don't mean that pure bank or a company that they'll take over your account system that would be bad what I do mean is that at some point there's some source of identity some credential that goes with that they're doing so on their own behalf and not only as an employee through an X500 director of LDAP service through their employer but not only in their capacity as a Google customer but actually they themselves can consent at some point that's important like we consent to be governed yet there's no mechanism for that consent is a two person, two or more party transaction, I can't consent to myself two different parties so consenting from inside somebody's account system actually does not it doesn't, it's illogical somehow people need to have an identity from which they identify themselves and can offer consent and can revoke consent meaningfully like sometimes through no fault of our own we don't revoke consent correctly like just hacking this app we frequently thought we were deactivating things and disconnecting endpoints and we hadn't being able to have people stand externally is important to keep everything honest and also for portability so the way we do that is through a JSON file so actually it was very patient with me and he instead of just having environmental conditions in Heroku or putting things in a Mongo databases you normally would for an account system and maybe having aspects of an account or a profile of a user just be dynamically generated from whatever you have laying around and not necessarily subject to a very high bar we just, we took a subset and said the account, the identity of the person is what exists in an account file and that file is a JSON object UTF-8 encoded just right on the server in a directory where we're putting them all what that means is that at any given point a person could look at their file their account and they could, you know JSON isn't particularly readable but it's not unreadable and it can certainly be arrayed so people could see oh here's my name, here's my current authorizations, here's my address here is my identity in a sense and what we're doing with the apps for this is we're actually having people go, for this service we're having them go to a couple of different apps one of them is Google, another one's a Node service we don't really want Google to own the identity of our people, we don't also want to run an entire identity service for these apps so what we're doing is we're encapsulating identity in a UTF-8 encoded file and guess what now you get portability so this is hard with identity systems to think of portability across domains it can be tricky there's a lot of ways to do it that are inefficient and hard to understand I can tell you firsthand but this may be a simpler way which is if you had a system of identity as consumers or citizens that was encased in an account file that represented your identity your core identity, your root identity and you want to go say from Google hosting it to doing it on your own service or some MIT, ACLU service, whatever, some identity RS service, you can port it you can move the file keep a system of pointers so people know where that file is and it can always have a non-authoritative source in more performance data driven systems but the authoritative source of the data the account file can literally be a hashed secure publicly verifiable file that you could read, know, have, own you can own your own identity and be a free person that is capable again of consenting to be governed for example you can and here's how so that's one of the features second thing is and it's working which is the most remarkable part so we forget how to make file, modify file because it gets buried in these frameworks and the complexity is hidden and so this is just like we're moving from Heroku to Ubuntu next on AWS in a mirror server in the media lab and I'm really thinking wow it's terrific go back to the primitives and the rudiments of just create file, modify file and not being lost in frameworks that disintermediate us from fundamentally what's happening with the authoritative data so for some things that's okay for a lot of things it's simple and it's faster for certain things it may be better to go to the old ways you know like a change mod and you know like an actual file that in that sense is authoritative here is an example of the how people add accounts and there we go there's the authoritative file and when you so right now I think she has done a really terrific job after some back and forth but she just represents them from the services that we currently have registered which is GitHub and Google you can see it's like true or false and so the true ones are the ones that are currently authorized we have precisely one scope for each service which is login the next step on the roadmap you'll hear in a moment will be to bring on a few more scopes like you know give me a copy of my account file so that's create a post it's basically can make a commit in my inner repository or you know put something in Drive for example the corresponding scopes we will bring on MIT shortly that will be a third and because we're really relying on OpenID Connect design for OAuth 2 it's not just OAuth 2 like that leaves most things to your imagination OpenID Connect is slightly more defined extension of OAuth 2 that includes a concept of certain endpoints and certain common claims for more enterprise-facing but it still works perfectly well for individuals so we're using OpenID Connect version of OAuth 2 totally backward compliant so any OAuth 2 thing that you guys are doing your clients are doing does not break when you add OpenID Connect what you do get however is for example a dynamic registration of services the same way across infinitely scalable networks and the ability to propagate types of registries needed to see in our case if for now we have say 50 providers that we allow and another OAuth provider has say 75 that is absolutely knowable at endpoints you know that can stay in sync in ways that I won't get into right now okay so let's go forward deactivate is critical you know consent is meaningless unless you can revoke consent just leave it right there and I'd say if there's three words that I would use for the kind of the design mantra it would be number one interoperability or almost like framework and technology agnosticism which is why it is we are utilizing you know things that you could implement the same way in python and Django or in node or elsewhere it's UTF-8 encoded JSON files the authoritative source you know we keep the authorizations in a certain directory that's expected few other things like that so I'd say that's for maximum interoperability but mostly agnosticism to whatever technologies and stacks people want to use that's really up to them number two is just more how if you configure a certain way you can get certain advantages engineering advantage or identity number two I'd say is extensibility and by that what I mean is if you want if people are adding on say in their agreements some login we get hub and Google today tomorrow there could be you know Schmugel or whatever some other thing that we don't know about now really by onboarding as you normally do in your business your contract terms can now keep pace by modularizing and encapsulating legal content we can now begin to engineer legal content so that it cannot be so chaotic feels like a quicksand sometimes some of this can be engineered through workflow so the extensibility allows you to basically extend the terms but also the parties just like snapping in a new service to a good service bus you don't have to re-engineer the service same with the loss it's extensible with the legal terms it's extensible with the technical service just because it's just OAuth 2 and the final one I'd say is portability so we've got by that I mean literally a user can now finally consider their the account file something that is theirs truly theirs that is merely hosted by a steward and that they can take from one provider to another so in our case it's literally coming up with a smart cities thing we're doing with Kansas City, Missouri and with some law schools and MIT and other places and we already have had some students go from a law school to a foundation that's funding and other places and it happens a lot in the world when people move from one job to another first day last day onboarding and offboarding is a nightmare I see dead people all the time on identity jobs there's like in NASA lots of places that's like not secret deep provisioning accounts is we don't do that right because we have not engineered it and part of that's legal you know frankly just take responsibility but now the web is not a fad it's time to engineer business legal and technical processes with a concept that core relationships need to be understandable they need to be trend visible and we need to be able to have levers for managing the relationships and that allow for accountability but more of the point allow people to make choices and put together new packages, new deals new value that's reliable that's legit I will just show you quickly here's the this is live demo of code still like dripping wet what will happen nobody knows I'm clicking the button nothing's happening okay something's happening what do you think should I authorize it okay I want to authorize it aha oh that's okay that's actually supposed to say that so let's go to my account page and haha just you know refresh that haha account page oh maybe I have to go home first that was user error by the way actually did it exactly right so here we have something this is literally looking for my file I've got google and not github at this point I asked him to put a login account which is coming from Heroku just so people can see while the page is definitely changing and it is changing like this is a real page got some stuff we pulled from the profile I can authorize github from here but more of the point look at this little guy here this is my favorite part I just click on this and I go oh at last I can see my briefing account for real I just click that and here is a file here is the path it's stable the name, location blah blah blah my current things are google true github false I think that's pretty good we can hash that we can check it up on the blockchain we can port it we can have it last with it and now let's go ahead and authorize I will authorize github so I'll just do it from inside the session because oh and I had already sort of given the permission before so it remembers that I had the permission I guess let me just click this again it should now say that I have both of them authorized I should have like completely logged out of everything in advance so you can see the permissions but now on the page here are the terms so I put some placeholder boilerplate stuff up here but this would be like you agree with apps are us that you own your data like reliable for nothing but just a little bit like a little boilerplate at the top and then right to the off terms that's the fundamental nature of the deal that's the sum total of the obligations of parties that is contract gaviolts talk in a moment about how we've encapsulated these into cards so you can basically legal components one clause per component chuck them in but here's googles like I've authorized Google to do this I've authorized github to do that then we'll have some back matter of the contract and disputes will be resolved in Bangladesh whatever so those will all be cards as well and let's see so let's now go to gosh 114 so let's go to Q and A I think and actually what am I forgetting actually I did not show the roadmap but would you like to say anything before I think actually it's actually at work right now he's from India there you go hello are you there so so number one thank you for coding this you know in volunteer spirit and anything you'd like to add yeah definitely so right now the prototype is you know on the server which runs runs node and you have simple html cs in javascript in front and the identity files are on the file system of the server so people could think of this server as a single point of failure right so in the roadmap we have things like integrating this on ipfs so ipfs is something hey Daza you can see how me yes I'm just looking for the slides sorry okay so by integrating id file on the ipfs what happens is the file is not a single system but it's on a peer to peer distributed file system so this way there's no single point of failure and though the server might fail the id file is still available on other file systems so this is one thing that we doing and the other thing would be having this dynamic contract on the ethereum blockchain so this way you know we're not limiting ourselves to the regular web but also moving this infrastructure on the blockchain which is much more reliable and better yeah thank you and so part of so I'm just pulling up this is the repo here we've got a a good whack of issues for something that we just started already but you can see most of what actually said there on the road map which is just to emphasize if you couldn't hear the signal that well we're going to be looking for distributed file sharing so that the server is not such a point of failure especially with just the research implementations we have now but also frankly at larger scale that's good for resilience it's good for trustworthiness it's good for reliability there's business models there we're going to be using ipfs it's a blockchain based distributed thing there's also something that Gabriella discovered at seasale called solid I guess that's another distributed something that Tim Berners-Lee invented and it's supposed to be amazing but we could use github for just distributing the authoritative account files but so distribution of the data is one thing and other things will be bringing on more scopes will be bringing on more identity and service providers so we can really see that dynamic contract Gabriella worked hard last night implementing bootstrap and some material design cards so we could decouple the front end better from what's happening in the back end and encapsulate those legal terms as true legal components so that the clauses and the authorizations are humming together in perfect synchronous harmony at last and what else on the roadmap some big things are on the roadmap we'll be implementing it so we kind of tore the identity system out of the app that we can use for so we can get it right in an OAuth pattern in time for this talk I think that's been done now we'll be reintegrating it with this app which is called Unworkshop it's basically just a way that everyone can do ideation at the same time brainstorm, break things actually if you sequence those the right way you can do some cool stuff so we'll be doing that next reintegrate with that app and then with a few more apps to make sure we have a identity as a service that's working appropriately good lord I think that's oh and then the lawyer's coming so in May we have what we're calling Algorithmic Law Conference which is also on the roadmap to invent Algorithmic Law shortly before then or at the conference no later than the conference we'll figure out how to better have data driven rules and dynamic systems for contracts maybe to get better analytics on what clauses to pop in which ones are litigated which are more performance better templates things that work you know you can do a lot of interesting analytics on contracts and so we're having a you're all welcome here to come to the Media Lab May 5th and 6th to an Algorithmic Law event where we'll basically battle test some of these ideas more fully and the last thing I'll say is that Gabriella and actually kindly agreed to show up at our favorite Tuesday half-night with Code for America Cambridge Innovation Center a week from today and we'll be hacking on I-Off so anybody that's interested in front-end Gabriella is a genius frankly and a hard worker but we can all use some help when it comes to encapsulating legal terms in material design cards Lord knows and so if anyone's into front-end stuff and wants to remake the law by encapsulating it and engineering it like normal people come on by also in the back end if you would like to the Node app and then our migration to Ubuntu which is important now that we prove the concept it doesn't really real until it's just on Ubuntu or Debian or something like that in the open and then we'll kind of dockerize different things but we have to migrate it or like I'll feel the baby isn't like safe it's not viable yet until it's just on a normal server with not like the special incubations of Heroku some of the things that will be happening next on Tuesday night just up the road one Broadway 7pm Code for America come on by so with that I'll hand the mic back to Hedy okay thank you we have time for two questions so who has a question here I don't I have one but some questions on faces yeah I think we can get the questions at the same time and then you guys here comes questions I got a github mug so just to summarize a big goal of this project is to put users in control of their identifying data that they share with third parties and make the contracts explicit and human readable yeah and so just for our friends that really tack this the gentleman confirmed that the main goal is to put users in control of their and then he said identifiable data and I'd say of their identity represented by more or less you can think of it like a title to a car that indicates ownership but that's right and then number two to make the legal terms more explicit and that's confirmed and confirmed alright so the question is what is the contract with those third parties after you revoke their access are you requiring them legally to believe what they have on and for you or the idea that the data is so dynamic that users are revoking continued access is that a problem you might address here great and the question was um what's the deal after revocation with the data on external systems are we seeking to control that as well like have it deleted for example and the answer is that's the question that we have to be able to answer at scale and right now it's a chaotic hot mess and it's molten hot it's unanswerable it's not tractable in my view the way things are now um my aim with this is to have it be better without changing anything on the terms or expectations of what people do with data because it's so much better just to use this it's so much more organized it's so much more reliable it's so much more efficient and effective makes new combinations and packaging a product and compositions of services possible that it sells itself and it tracks open source people and other people to add to it and then the master plan is a new deal a very old deal on data basically the deal of America and the commonwealth is that we're free people and free markets I took a note to defend the constitution of Massachusetts when I was a lawyer and it says pretty plainly it's like on our website I put it there in 97 it says the deal is that like we're consent to be governed so the real deal is that we can revoke consent and that would mean in some cases that if the consent was to have data and revoking it meant in our terms that you delete the data you delete the data if what it meant is that you didn't delete the data but you don't get further data we would see that in the card and when we revoked it that would be an auditable locked thing so the plan is to create enforceable engineered basis that people can know the deal and that I think a competitive market for people that would be willing to delete your data and to say that they'll do that no off permissions would emerge and that if some didn't do it for reasons of incompetence, error evilness or whatever we can know that and now we have something you can easily almost machinably bring into court for enforcement you know we have to go to court like you can have liquidated damages like okay 50 bucks in PayPal if you didn't delete you know these five things and then look at the span naughty so like now we have enforceable engineered scalable basis to address ownership of data and deletion is the best way to know who owns something they can destroy it they totally own it so I think you've asked the right question what was the second question you're welcome what's your name Tyson okay next good question what's the second one cool the second question is that you mentioned a few people that you're already going to work do you want to find it what makes them so excited about this project and how do you find it thanks so we have a bunch of like what I would call identity at MIT and then she crossed America like mostly in the west coast but New York and a bunch of places some terrific companies RSA here but many like are true innovators of identity services and they start work for what we call a user centered perspective so the center of all that goodness is something called Internet Identity Workshop meets twice it's an unworked conference that meets twice a year in Mountain View and it's where like all the people really really into this like below IETF below W3C and below like the trade shows like that's where we really get together and kind of have it out have a great time and event things like OpenID Connect came from there OAuth2 came from there from one of the sessions like many great things came from their super upstream we're inherently interested in identity those people and so like we kind of fuel each other I run something in that group called the legal forum we've been getting good feedback from people and also some other standards groups where we've got initial feedback so I found them just really just the people I currently hang out with that have been validating it the real power though is with Akshith and with Gabriella and with the people that you can't see who are labored is volunteers that they saw there's an opportunity to do something better and they basically volunteered civic hackers to contribute one way or another to this and you know luckily it's unknowable because thanks to GitHub we have you know commits and we have issues that go way back across some repositories that we'll be linking but it was if Akshith and Gabriella had not when we were finishing up a blockchain thing expressed interest and even passion about this and a sense that they could do it and that it was important to do and volunteered to do it late at night when you know Gabriella has a great boyfriend that she like wasn't with him throughout time and she's got a lot of money on this and then Akshith has got a job and a bunch of stuff to do like we've been volunteering on this for a while their passion and willingness to volunteer that indicated to me that the time is right it wasn't ready two years ago I don't think anyone even understood what I was talking about people understand it and people are ready to help and so I recognize that pattern pretty well it's something that's time is right so we're going for it also you come to think of it so like pivotal labs a tech talk first I was going to do a boring buzzword laden thing because I thought that was what I was supposed to do I wasn't even looking forward to it that much although I wanted to see what was going on in here so like pivotal tracker and so but the interest in is it shine fun when I just said well I guess we could do how about IOS because it's been in the back of my head lately so due to blockchain things were coming up that suggested might be time maybe and so I just trotted out there and she's like her eyes lit up just like yeah like you know our clients have that as an issue we don't really you know there's not a good design pattern it's interesting to flip it flip the you know pyramid and put the user at the center so she validated it so I'd say literally if it hadn't been for pivotal labs catalyzing the topic with the talk which then required like some kind of demo this wouldn't be happening so I'd say the main characters volunteers MIT people and you thank you so much thank you so Gabriella and actually thank you for your being here and for everything you've done and you can hang out with all of us and hack this into true existence a week from today at Code for America and we'll put a meet up on Massachusetts legal hackers in a Code for America to circulate around to people here I'd be grateful would love to see you again I'll stick around for a while for a conversation and with that I thank you very much and we're going off here