Private Social Graphs: Trust, Consent, and Access for Real Communities

Transcript

Hello everyone. Uh, nice to see some friendly Frances here. So today I'm going to talk about private social graphs, trust, consent, and access for real communities. And when I say real communities, I mean humans. So before we we start, let's define very quickly what is trust. Anyone here know what trust is? All right. So trust involves a willingness to rely on the action, words or decisions of another party with the expectation that they will behave in a reliable, honest and competent manner.

Um and trust is really needed when it comes to uh us to want if we want to transport ourselves from a known place to an unknown place and there's a lot of uncertainty. We need trust in order to be able to move from one place to another. So in reality trust is the lubricant for society. So what are the core drivers of trust? The core driver drivers of trust um are authenticity, empathy and competence. So essentially if we have all of this we can say that we can rely on on someone.

So how can we increase trust? We humans we have had many different protocols for establishing trust. Uh we have the handshake protocol, we have clinking glasses. Uh for you guys that don't know um we started using clinking grasses uh back in the back let's say when the monarchs they were afraid of being poisoned they were clinking grasses and making sure that the glass from from their cup will spell also the other people's uh cup. So like that if the if the king was going to be let's say uh poison the other people will also be poisoned. So that's how we started actually like having this protocol of clinking glasses just to establish trust

and then we have like pinky promises like if we do like this with our friends we're like hey please trust me that I'm going to actually do what I'm saying that I'm going to do and there are a lot of human protocols that have been developed uh through humanity um and and they've been changing um from the let's say from the physical space to the digital space And uh there is a really cool game um if you guys scan the QR code, you can play play it out later. Uh it talks about how uh trust uh can be built and we can have let's say um uh a way for us to create the perfect environment to start to

let's say increase the trust that we have for each other. So for us to increase the trust, we need to have repetitive interactions that generate possible win-wins and hopefully also that have low miscommunication. So like that people can let's say rely on each other's words. So yeah, as I said before, trust has evolved over time. We used to be let's say um so societies or tribes that were uh relying on local trusts. Then when the societies those societies got too big, we started having institutions that we delegated the we delegated them the power to decide um who are the people

that uh let's say they want to back up with their when it comes to university degrees or when it comes to land uh registry or any other forms. um um of institutional way to say hey trust what we're saying and now we're moving from the local to the institutional to a distributed system of trust and where I think is uh is very exciting for humanity and for for us right now that we're seeing that there's more and more uh AI and so on we need to go back actually to the person human to human but yet we can create distributed distributed trust TR.

So um as I said before trust has been evolving quite a lot. Um so trust um also travels trust has carriers and yeah a w a witness let's say before we were like a local local tribes and we would rely on what the local people would say or what the people within the community would say. But then we started okay if if we if the commun if the people are in different parts of the world or in different let's say locations how how will the trust be transmitted? Well, the trust was transmitted with witnesses or let's say or with seals um as you can see like for example back in the days

you the king will pull at seal um and yeah or credentials or signatures that people recognize and if they recognize and if they actually think that they're legit and true then they will trust the carrier and yeah so For trust to be able to exist, it has to be a shared memory. And um yeah, within the communities, the share memory is it used to be uh personto person just with um protocol that we call the gossip protocol, which was basically people talking to each other um and telling each other like

what they believe and who they trusted. And yeah usually they this share memory will be created when people meet in real life meet in person. Um but then you know when we started developing writing and we we started developing other forms of accounting the memory could be kept um and the memory could be let's say um transmitted to the pe to the people that were close to the let's say to the witness area meaning the people that will receive any of the seals or receive any of the witnesses that they will come and share that context with them.

Um in in the in the age when pirates uh let's say were a big thing um the rule the the sea had a different forms of rules. Each pirate ship had their own rules um and yeah the the since the crew they couldn't let's say have an arbiter they couldn't have someone um that will say hey what this person is saying is right and what this person is saying is wrong. um they actually had to coordinate themselves because they were in the ocean and they couldn't rely on hey let's call the let's call the core in Spain they didn't have that so they had to have let's say um um a way to govern themselves within

the crew and yeah in the crew articles you see um that many of these different type of let's say pirate ships they will have uh how the crew divided the shares between them. Um what what are the different type of principles that they have to follow when it comes to conduct? Um what is the compens compensation after let's say they got some loot or they got let's say rewarded by someone. Um and what are the decisions that belong to each crew member and yeah the authority in these let's say setup in uh in in in the pirate ships it was contextual. So the captain

will led for example battles. The quarter master would would handle all the shares and if there were disputes between the people they will this person will be in charge of that and there were other people that had different roles and they were delegated trusts to actually uh fulfill those roles. So um yeah, as I said before, there were many different ways for people to um transmit or carry um let's say different forms of stress. One of them is a wax seal that you know it was used in order to let's say prove okay that uh that this happened in this moment because I had the stamp to do it um which was not

cheap to replicate. So that's why at that time it made sense. Maybe right now it's too easy to replicate that. So right now maybe the wax seal makes no sense. um the matching halves. They will come and they will put let's say a stick and they will they will uh start accounting uh what are the different depths and then that that uh stick they will divide it in two and then each person the one the the one that owns the debt and the ones that that the debt is on to they will have these sticks and they and that will represent like what is the debt based on the accounting that they will put on these sticks and they will have a matching thing. So when they come together they'll be like yes this

is what we agreed on. Um so it was a way to not just leave it to the memory uh to the context of our memory but they had this context that they could bring together and they could match it and see and verify hey actually this is how we do it. Um then now when we start inventing money there's different people that started using the serial numbers as a way of recognizing each other. Like I would say uh to my friend Nate, "Hey brother, I gave this uh um Venezuelan bolivar with this serial number to this friend of mine. If you meet him and if this person has a serial number, please treat him as my brother because he's

been treating me great and I don't know I've been sleeping in their house for example." Uh this is just a clear example of how like let's say trust can be um delegated or delegated. you know let's say uh carry using different different type of instruments or different type of tools um then as I said before when we pass from uh having this local um trust and the community trust we started having let's say issuers that needed to um prove something so we had government courts um professional bodies that were basically establishing what is it that we need to trust. So um the university credentials basically attests that someone went

through their university program and that hopefully they pass all the exams without cheating and that actually they are capable of doing um what the program supposed to be for. Um the ownership records use documents that are issued uh by land registries or by uh governments uh in order to let's say in order to uh specify who are the different owners and what are the different conditions that they have. Um so yeah as I said before um the idea is that we pass from uh an institutional uh type of trust to a distributed type

of truss. The institutional type of trust is very opaque is closed centralized is licensed is top down. What the distributed type of trust could be transparent if we wanted to be transparent or we could have let's say um um we can share some information with people that we want to share. So we can have selective disclosure uh we can have we the distributed um form of trust can be inclusive decentralized account and accountable and also bottom up. But yeah, the the idea of having the say a form of distributed uh trust um is possible but we are very far away from it. Um why? because we still haven't

come as a society and as humanity and we haven't created new protocols that can enable us to carry this trust um across across the time and then across all the different uh let's say um new mediums that we're using whether they're digital or physical. And yeah, essentially um people usually make decisions using uh overlapping trust claims. So um it could be let's say a trust claim could be a document but then it could be my friend that witness it or it could be uh let's say um a local record or a history or an institutional credential.

And based on all of those trust uh based on all of those I'm sorry uh proofs humans the way that we act normally is that we decide if we're going to trust it or not. Uh that's that's how humans work. Let's say if my mom tells me today, hey manu go and meet this person and this person let's say is not necessarily the best. I'd be like in my in my mind my mom is losing some trust. It's not that my it's not that trust go to zero to 100. It's not it's not a point system, but it's losing some trust in my mind. But again, the trust is very contextual. If my mom tells me again, go meet another person and the person is again not necessarily the best

one, then my mom is losing let's say trust uh when it comes to that specific context. But when it comes to my mom telling me, "Hey Manu, go and check this restaurant." Damn, my mom is really good at that. So, I might trust my mom on that. So, all what I'm saying is like the trust is very contextual. Trust is not a zero or one. And trust cannot be a general number either because it really depends on what we're talking about. We might trust someone uh for buying, let's say, a Pokemon card because they have a lot of knowledge when it comes to Pokemon, but we you might not want that same person. you might not want to trust

them to buy, let's say, I don't know, a computer for you because maybe they have no idea about computers. So, like trust really, really, really depends on what we're talking about. Um in the age of the internet plat platforms can capture distrust and they can gatekeep it and they can actually uh conflate the um the let's say the the different things that are happening within within the different connections that are happening within let's say communities the platforms they can literally if they want they can misinterpretate everything or or make it look in a different in a way that the platforms want. So the platforms have a huge um um let's say

opportunity to capture uh what is being communicated and the trust that is being delegated within uh communities in the age uh let's not talk about pirates anymore. Let's talk about cipher punks. In the cipher the cipher punks they start having gatherings and the gatherings they use it to exchange um uh exchange uh keys. So they could communicate with each other without without having governments or without having any not even the platforms uh intercepting the communication that they had. So the cipher response what they did is that they started to have gatherings where they would come together and they would exchange the

public keys and they would keep the private keys to themselves so they could start communicating with people that they validated and vouched for in real life. So that's why actually the the um um key party ceremonies um by the cipher punks were were happening and and are still happening today. So the idea is that you have a key signing party where you meet face to face. You check the person in front of you. You conf you you compare the uh the keys that they're giving you. You verify it's actually this person is holding this key. And then later from your own device you can actually use that key and start

communicating directly uh to those individuals. And there's other forms of uh uh creating share secrets. Um before let's say if we uh let's say let's not think about digitally but let's think about physically. Let's say we come together and we break a a base um ceramic base and we're I don't know 10 people here. We break that ceramic base and hopefully we break it into 10 parts and those 10 parts if we come together and we are able to put put it together it means that you were there before.

So I'm saying this just for you guys to have an idea that these type of concepts has been used before. What we're doing is a remix right now. We're using share secrets um and Adish Shamir basically show how basic different people have can come with different parts of the puzzle and that can reveal secrets. And why is this important is because the share secrets um um can be context that can be shared with all the with all the people that are uh um let's say are supposed to be seeing that share secret.

So what are the ceremonies? So why does ceremony scary? Um the idea is that these ceremonies people um people know who are the people that are taking part of it. Um, everyone can compare the all all the different private keys. Um, and also they can create a ceremony where they say all of these different people were there. And yeah, if you don't want to participate, you also don't need to participate. It's not necessarily force. So my um um analysis is that gatherings can be used as an infrastructure for generating trust like coming here to Denver to Denver no I'm sorry to Mumbai

um is uh is um um let's say an an incredible opportunity for us to create a to create let's say the right environment for trust. Um so I'm right now speaking here because someone trusted that I might have the knowledge to be here speaking to speaking to you guys uh to you guys today. Um and then some of you might be here so because someone trusted that you should be uh that you should be here at Defcon and gave you a ticket to be here. So like if we think about it uh it's all related to trust. Trust is actually in my opinion the the ultimate currency. uh and in this world of AI and so on trust I

believe is going to become way more important. So what I'm thinking is that we can start creating new groups and new forms of trust where we can have consentbased private social graphs and I believe this is the the key and uh the key for freedom um is that we can is have these social graphs that can share trust uh in different context. Maybe let's say I want to share with this group of people this type of context that is relevant to them. Maybe I want to share with my best friends only who are the pe who are all the people that have scammed me so they are aware of it.

Maybe I don't want to share that publicly because that can put a target in my head because I'm doing that. So maybe I share it only privately. So imagine that we can start creating different forms of groups where uh trust and context is being shared privately and we can have selective disclosure of what is being shared. So we have been experimenting with uh vouching vouching systems. So who is making the claim uh is important. Um can can we rely on the claim that they're making? um which decision needs needs the claim and who is actually able to see all of these claims.

So um I'm going to pass the other one just because I have very uh limited amount of time. Um so yeah, we can also create expiring claims. So let's say if you don't uh if we create a protocol where we say hey we're going to meet every new moon or we're going to meet every full moon and if the people don't show up every new moon every full moon or new moon or whatever we decide then they can their their trust can start to decay for example we can create that type of systems where the the the trust could decay or could expire completely. Um so we can create new forms of let's say protocols for us to share context and

share trust and it's important to have local stewards. Why? Because those local stewards they are actually like bookkeepers. They keep that they keep that trust. And when I say local stores, I don't mean necessarily like physical local um um um people, but I mean someone within the community that could that could be let's say name store and explain the context of all the decisions that were made because currently in the society that we have there are many decisions that there are many laws that we don't even know what the context uh why these laws even exist. Um, so it's important to sh to whenever we make let's say a decision we have that we that we're able to capture the context

and all the people that were involved. So like that we can come back to it and be like oh this is why we made this decision like our soci society has been losing all of that. So questions that we should be asking before we make this new protocol that I'm inviting you guys to think about is what must be public, what should stay local, who can make a claim, how can someone withdraw a consent, when should the claim expire, how can a decision be challenged, and there's way more other questions that we should be asking when it comes to how we create this new protocol that will enable us to uh um let's say exchange different things. And uh we don't need to create a marketplace. we can create literally a place where we can exchange any form of

um services, any form of let's say um tools that we might have um or context just share context. So yeah um if you guys want to talk a little bit more about um private graphs and trust and consent um please uh follow me, talk to me on Twitter, Telegram or whatever. And here you have a QR code where um you can find all the videos that I that I made uh in the last three years about these type of things.

And yeah, thank you so much. >> All right, thank you. Thank you so much. Well, we have some questions. Um well, let's go with the first one. How do you handle low miscommunication when consent is private and can be added by others? Um, first of all, the the uh who who made that question? Anyone here made the question? >> Could be coming from online. >> Okay. Okay. Um, so yeah, I'm first of all I'm wondering I'm wondering what the problem is because the low miscommunication can happen for sure but the when it's private if it's between the two

individuals uh how can you handle miscommunication by sharing feelings and by being primal just sharing just literally sharing um your opinions. Um low low miscommunication low miscommunication uh create creates the opportunity for us to have conflicts. >> All right. And uh how can we then verify the authenticity, origin and accuracy of information within private social groups without compromising user privacy or consent?

>> Yeah, this uh this actually um this is a good one. So essentially it's like who who are the different type of people that we trust for different type of topics. So there might be someone that is making a claim about something that happened here in Mumbai but they're maybe not even here in Mumbai. So do they have the context? Do they have like do they have did they experience it? Did they leave it? So you as a the my idea is that rather than having a trust score that is given by a machine each individual will make the will score

everything contextually based on who is giving you that information. What type of let's say uh live experience they have had what type of context do they have? Uh what type of connections do they have to that specific manner that they're claiming. So we humans already do that when let's say we connect on different social medias and we and I see I have 90 uh uh um common let's say followers or um that means nothing that means nothing what is the context of those followers what is the how do how how did we meet them um that I'm just giving that as an example because again the most important part is like who is actually giving that information and that information can be

sent privately And you can you can let's say say hey my friend uh my friend is telling me that this happened to him and he can prove it because he was there he saw it blah blah blah and then it's on me to say I accept or are not accept this claim that this friend is making so yeah that's how we humans already work we just need I think we're just going to digitize this type of thing in the in in the near future but but yeah we humans already work like that.

>> All right, thank you so much. Now we are ran out of time. So, apologies about the one minute.