Privacy in Other Words, Infrastructure for Expressive Freedom

Transcript

Good morning everyone. Thanks all for joining us here today. I am here um to be a narrative. Um Jessica and Eric Alston wrote this paper that we'll be discussing it today. It's called privacy. In other words, infrastructures for expressive freedom. And um I'm going to give you a very brief overview of the paper itself. Um after which we'll have an a discussion by uh Puja and Anush. So let's get into it. Um so Jesse Cuts and Eric Alston um wrote this paper and um Jessica is working for an NGO called speech without

borders which uh wrote this paper independently. Um they got feedback from the first amendment lawyer Preston Burn and uh were inspired by Denny Ryan who at the last def connect and Buenosidis made this argument that um uh people in finance and institutions don't often care about the properties that Ethereum brings about until you were frame it as eliminating counterparty risk. And uh in this paper, Jessica Cuts and Eric Ellson uh attempt to make a similar argument applying it to privacy.

So privacy in other words, um that's what the the title is really what the paper is all about. And um when people don't value privacy on its own terms, then they translate it into a right that they already value and cherish. in the United States um that right is the freedom of expression and freedom of association and that is the core of their papers argument. Now, now why looking at the United States? Because that is because the United States are quite different from many other countries around the world. Uh in the G G20, I believe the US is the only one without a fundamental law for privacy. I might be wrong. I'm not an expert here,

but there is certainly a gap that they're trying to fill argumentatively. Um so, so that means what you disclose publicly has little protection by default. And of of course public blockchains like Ethereum are open by design and they should be. Um so in the US um the strongest footing for onchain privacy is free freedom of expression as the authors argue. Um why looking at the US specifically and why does this matter beyond the US?

We're here in Mumbai today. Many of us uh don't reside in the US. So so why should this concern us? Um, so I think the link underneath that is is profoundly uh human. We're not talking about uh US law specifics necessarily, but but privacy as a fundamental human right around the world across cultures. And um we can we can learn we can see from our own lift experiences perhaps that when people know that they're being watched that they may stay away from meeting from associations from from donations from voting from any act that may reveal information or context that they deem private. Um and also because

infrastructure is global of course uh a lot of infrastructure built or or serviced through US institutions or companies is it's global and we may use it every day. So there's another link here. Um so coming back to the paper central example um in Ethereum that can mean many things. So looking for example at a funding round where you make contributions um they can double as votes. So in the DAO space you may be familiar with a variety of voting mechanisms where it's not just about giving uh you know contributing to public good funding

itself but it doubles as uh voicing uh an an opinion making a vote. Um, another example could be perhaps um the expressive part from the financial part um where you would be um examining every transaction for its purpose and if you do that you you would reveal um private information. Um here the paper argues that privacy should be the default because on a transparent ledger choosing privacy is itself a signal.

Right? You can see that um with the the public nature of the chain if you opt in to some process that preserves privacy but the default does not ensure these privacy guarantees. then the act of opting it in is often visible through directly through the act or through through some metadata. Right? So um I think what is to be taken away from this I think um I would really encourage you to read the paper afterwards if you're interested but the the fundamental high level takeaway I think is that privacy technology is infrastructure for expressive freedom as

the authors argue and in other words on an open ledger privacy is what makes freedom of expression possible. You can find this paper on their website which is speech.ng/privacy minus in minus other minus words. So that's speech.go and uh in the part one you have a short overview uh and that's definitely a good read if you don't want to read the full paper.

Let me close with the paper's last line. Um, and I'll quote this uh, literally verbally. In the meantime, the Ethereum ecosystem is building technologies that give people the option to exercise rights they already have within rules that already exist. So, there's a lot to unpack here. Let's hand it over to our two panelists and speakers. Puja a plurality thinker and she's affiliated with a getting plurality group at the Harvard Howard Allen lab and Anush uh he's a computer science and

philosophy researcher and he's building a payflip XYZ a privacy forward payment protocol. Right Pua over to you. >> Sure. So um just for context so Jesse who is the author of the paper couldn't be here uh for visa issues and we were supposed to have something of a a debate or conversation. So um I will try to steal man the argument but I I also want to give a slightly different perspective. So what I agree with the intuition and I think the intuition we all have here is that if you have surveillance right that chills association that chills speech and so there is a general reaction well you need privacy. Now that intuition is I agree with right 100%. I'm a privacy

maxi. I like encryption. But I also think that uh privacy is a richer concept than I think what the paper implies. And you know it the paper actually doesn't provide a definition of privacy. Um it's it says privacy as default. But what the implication is is that privacy is uh a unilateral right to disclose information when you want to whom you want. Um and so it's you know I get to reveal what I want whenever I want to whom I want. But I actually don't think that is faithful to notions of privacy that we have and it's underinclusive and overinclusive um in ways that actually uh make it hard

to check power. So let me give you a different definition of privacy and then walk through some quick examples. Um so my notion of privacy is privacy as contextual integrity. So I think of privacy as a bundle of decomposible rights on the uses, abuses and profits over information that we generate in social context. And when I say generate in social context, I mean in conversations with people. And so each context has different expectations of privacy. And so there isn't really a universal sort of rule or blanket, right, principle. It's really what you and I kind of agree to when we are

talking and generating information together. And so you know this matters when we're thinking about power. So for example, a sitting president of a nation state does not have the same expectations privacy that a citizen does. A president is an agent of the principle of the citizens, right? And they have a duty, right, to uh steward resources towards the public good. And so we expect, for example, the president of a nation state to disclose their tax returns and to disclose information, right, on their financial transactions to make sure they're not looting public goods, right? So there is no voluntary I'm going to disclose what

I want to whom whatever I want. It's actually a mandatory disclosure and it's part of the bargain of when we give power and we delegate power in this principal agent relationship. Now there are other principal agent relationships that we have even in running companies. When we give our capital to a manager or an executive or when a you know a company wants to be on public markets and get access to public capital, their expectations around financial disclosures and transactions that are mandatory and it's part of the bargain because shareholders are giving their hard-earned money, right, to an executive and they want to make sure the executive is not going to loot them or compromise maximizing shareholder value, right, to their own benefit. And this is why we have insider trading laws and

financial disclosures. So those are like two examples of where privacy that those are actually examples of privacy expectations where information has to be disclosed by somebody who's been delegated power right. Um and this is important right not just for the provisioning of public goods and private goods but also you know keeping adverse selection out of markets and and then there's also other notions of privacy like a nooge if you and I are talking to each other right and I you know surreptitiously record our conversation and sell it to a data broker I'm violating our expectations of privacy right but just because I do that in a shielded transaction and I voluntarily did it right um doesn't you

know and and and through an encrypted protocol doesn't mean that it was actually a legitimate it might might have been formally private but substantively I've actually undermined our privacy expectations and you know so similarly like you know when we're thinking about like since we're you know this is Ethereum conference and we're talking about encryption and blockchains you know a sitting president or an executive that does like shielded transactions for example launching a memecoin that enables to accept bribes right and like loot public goods right on the side or say, you know, betting against your company because you have insider knowledge and like profiting off a decline, right? Those are activities where they're sort of unilaterally and like serreptitiously, right? Um

disclosing information protected, right, through encryption, but those are clearly cases where they're violating privacy and not just privacy duties, right, for disclosure. So that would be sort of my nuance ad to the paper that um there needs to be some richer notion of privacy. Um and if not then then then a concept that's just purely tied to expression and not tied to the expectations that people have when they disclose information and the agreements that they have. Um then then you're actually what you're doing is you're not you know protecting the weak from the powerful. you're actually empower empowering the powerful to exploit you know exploit information asymmetries and

and you know privacy is something that's not uh you know you know a uniform rule for everybody in every situation. It's actually uh uh you know a general rule is you know the p the powerful have more to disclose and have less privacy expectations than the weak um you know to whom they own duties. >> Yeah. Um I think I'll add a few perspectives. So one is that um we have an idea in computer science which can be applied to privacy which is the idea of side effects that when you run any application let's say I am running an application where I'm changing somebody's date of birth is incorrectly

put right the idea of side effects is that by changing date of birth only the age should be changed not the name for example it's a very simple concept but it has like incredible ramifications when you build applications cuz when you want to make a change, you want to limit the change to exactly what the change should incur upon and the exact same thing when it comes to communication like what you said the insider training like if say in a case where you work in a company and I have shares in the company I get to know you you tell me on signal that we are going to do this updates you know it going to be incredible in the market and I buy more shares Right.

What's the problem with that? The problem with that is that what you tell me the effects of that are not limited just to that channel. There are side effects that go outside of that. It leads to market manipulation. So the idea is that many people argue and and there's a reason to it like why would I have to let the state the public or anybody else know what you and I are conversing which is mostly true as you said contextual integrity. But when in the context when you're communicating about something the effects of it can leak outside that very context can affect other people in ways that they didn't sign up to. Then of course you cannot just say I need to not reveal what I'm doing. So what is interesting

is this actually makes privacy all that more important because if we can provide privacy tech where we absolutely guarantee privacy where we can also provide guarantees that side effects underinsured then instead of the conversation of like oh like I like the example you gave like I can't even reveal if I'm sending this message for a vote or for a financial transaction and then I have the permission okay for a vote you have privacy for financial you don't cuz that's only leak of an information right but if rather these two things the only thing that I need to prove to the

protocol to the world to the public to stakeholders is that there will be no side effects outside of just I'm talking about like you are recording our conversation as long as I am aware of that that's not enough as long as I know no side effects like you will never reveal that to outside people even if you agreed that you will record it I'm fine with the recording so this is why I think what is important is this being able to bound the scope of any kind of a communication within the context of the communication this binding this this side effect prevention that is at key in every uh convers privacy and I think so much of time people ask and people demand like

uh we need to have KYC because we need to know if you're spending your money in like ways that you're not supposed to. Again the same side effects like if I'm sending you money and if within the context of what you and I agreed you spend the money within that and doesn't have any side effects that goes to the market to the world to the like you're funding a terrorist organization that's a huge side effect if we can absolutely certain that it remains within the binds of whatever we agreed and whatever is agreed upon by the protocol society the law the the state what what have you that's already I think one way to look at it in a way that does not demand reveal or not that kind of way,

>> right? And so like I think the the way to kind of like split the difference here is I think modern sort of like transaction encryption protocols need to find better ways to express social context. So one way was through credentials. Um I think there are more sophisticated ways now um that will come online but like to to your point Anuj like ju just just to give you a counterpoint like I actually think privacy as a concept is indexed to power. So the the paper makes the argument that like vote voting is an example of association which should be protected under the first amendment. But actually you know

votes it depends on who's voting like in the US constitution if you're a public official if you're in Congress and voting those votes have to be public because right you are an agent right of the principle of US citizens. And so US citizens to hold your congressmen accountable you need to know how they vote. But citizens have no duty of disclosure there on how they vote, right? And so that's a perfect example of actually where you have the same activity with completely different privacy expectations depending on what your position is. You know, whether you're a principal, whether you're an agent, right, whether you have power or not. And and you know, I would actually say like a lot of the conversation about privacy like when when a limitation is

people tend to focus on the f first amendment and fourth amendment. Um but actually I would argue that the US Constitution like writ large is actually a gigantic privacy document because what it's doing is you know across various provisions is scoping what information needs to be disclosed like you know even you know an argument that could have been made in the paper that wasn't was like uh protecting the inner conscience right that's like from the first amendment establishment clause fifth amendment where you don't have to testify yourself right that actually is a protected space of privacy. You can never compel somebody's mind. You can regulate action but not compel their mind. There's other parts of the constitution, right, that force um you

know public disclosure of of of the government. So public hearings, public trials, right? Those are actually privacy regulations saying you can't actually withhold that information. You have to disclose it for accountability. Um, so I I think when you like, you know, the constitution is actually kind of an attempt to to express privacy rules across various different contexts. Um, and and I think that's a a more a more useful frame and a more faithful frame. >> I mean when you say like the example that you gave uh the first one about like who votes is important, right?

Again comes back to like in many ways what I was saying that the side effects of a somebody high up in the institutions of power their vote has versus my vote who is a nobody for example and is very different like it stays within the context of like me supporting so and so candidate but whereas somebody who holds the power like a senator for example their vote matters a lot because they end up you know being part of the legislative body. Absolutely. And in that case when you do talk about conscience something like I've always been curious and this is a complete different uh perspective. So allow me to do that because if you think of privacy like the first time all of us humans realize privacy or learn about

privacy is as a babies when we are taught to hide our privates like you don't expose yourself outside like you're taught toilet training all that XYZ right forgive me this is this is crude what I'm saying >> very old testament reading >> it's oldest it's about shame I mean but it's true like it is about shame and this sense of shame and privacy somehow never goes away which is unfortunate because people keep thinking about privacy in terms of what I don't want to reveal what I want to hide and then unfortunately people who criticize privacy they say what do you have to hide and we need to move away from the hiding way of thinking of privacy and what are what it enables within what context are you actually doing it what enables who enables what powers are in

at play >> so to extend that religious analogy the new testament is about communion right and community right and and that's the shift away from shame right to I mean you you you can actually so nu is a theologian you can actually uh explain what the you know the paradigm shift is better than I can but that's exactly a situation where you know by and giving groups boundaries and privacy right around their context that you actually can enable a positive vision right of collaboration >> true >> so I think that opens up a whole new tangent which should be another panel or maybe a book but I I suggest We go to the mir card. We have uh questions.

Thank you for asking. Um seems like all three have one vote. So I'll just take the first one here. Uh so that question is do you think privacy can truly exist in crypto when governments are making transactions t taxable and requiring identity verification? How can we protect privacy while com complying with these regulations? I mean I think um again the question is not it truly exists or not but rather it's a matter of like can we um can we open can we have communities we are instead of revealing your identity it's rather revealed like what you are doing within the which within what bounds it's

happening. So currently the way we do identity verification no it absolutely has no like the level of disclosure you have to do is too much but rather if the disclosure could be more contextually driven like what is the transaction all about and based on which you actually reveal what needs to be done and that what it is all about is not actually evaluated by a person sitting in a government office or sitting even in a protocol office for that matter but rather if it is algorithmically possible then that would be to a certain extent better than what we have. But >> so yeah, so like there's an idea that we uh introduced my collaborators and I have this uh called community recovery where so you have a wallet and so say

there's some suspicion of money laundering or some illicit activity. Uh how do you actually right confirm the elicit activity? the bar for, you know, opening a wallet and not opening all of it, just opening the aspects that are relevant to the illicit activity would require uh a qualified majority of people who you are in communication with, but they're not in communication with each other. So that limits collusion. That's sort of the theoretical aspiration. And you know a qualified majority of them basically end up being a kind of like jury and a sort of check on what uh you know is this is there a you know

probable cause to search um and so that that I think is like a compromise where you can you know continue to have privacy in wallets right and to the extent that there is a suspicion right the suspicion is not right kicked to um a a a sort a centralized authority that can like force your wallet open but actually requires the set of people who that elicit activity is relevant to right and a diversified jury amongst them right that are talking to you but not talking to each other so they have enough local context so that's one way to strike that balance it's a theoretical idea that you know I've been working on quite a bit >> well I'm afraid that's all the time we

have um there's a lot more to discuss u my cury to read a paper at speech.go and uh yeah definitely thank you to Jesse Katz and Eric Alson for providing us with this fantastic resource. Unfortunately neither of them could be here today with us but I'm so glad that Hua and