Simon Brown: Welcome Carmit and Muthu. Please tell us a little bit about yourselves and your work.
Carmit Hazay, Co-Founder & Chief Cryptographer:
I've been working on secure multi party computation for the past 15 years, both practical and theoretical problems. Muthu and I have collaborated for the past six years, and he is my professional other half.
Muthu Venkitasubramaniam, Co-Founder & CTO:
Yes, that’s true. Even though we are geographically apart we do most of our collaborations over Zoom. I am an associate professor at University of Rochester. The focus of my research is cryptography, secure multi party computation, zero knowledge, both theoretical and practical aspects.
Our interest as a team is to design scalable secure multi party computation protocols, and I mean this in the general sense of scalability. Secure multi party computation encompasses things like zero knowledge, in which we are also interested. We’re market focused, and we are looking at what kind of multi-party computation and zero knowledge applications the market wants now. We are trying to build something that will be usable, that will be scalable, which is a pain point, currently.
Right now, we are focusing on financial Institutions and financial applications. Multi-party computation is an ocean - the applications are just an ocean. We want to start somewhere where we can bring these tools and techniques to the real world and get people to understand what they can do and what they mean. Once it gets to a point where people really understand the potential, it's going to be a revolution.
SB: You said you want to address some of the pain points that people are experiencing and the potential of this technology. What do you see as being the main benefits to people in general?
Muthu: I would begin by making a broader statement. Secure multi party computation and zero knowledge are advanced cryptographic techniques, and they encompass a wide range of cryptographic tasks. I would say that in the theory world, for example, if I wanted to prove any kind of measure, like complexity measure, I would go straight to secure multi party computation, because if I prove it for that, I prove it for everything, right? It is kind of the standard bearer of modern cryptography. So, let me give a little bit of an introduction to zero knowledge and multi-party computation.
In general, when people consider cryptographic techniques, they understand security is about securing data, about securing data at rest. Encryption is ubiquitous. You see this lock icon on your browser, and you understand things are going to be encrypted. Even people who don't understand the real definition of encryption understand the semantics. They understand that the data is just being transmitted from one end to another, that there isn't much computation going on, that it is data at rest. With zero knowledge and secure multi party computation, (zero knowledge is also more in this realm), it’s not just about data at rest. You also want to make use of data while guaranteeing privacy.
In a very broad sense, MPC helps people collaborate in a positive way. It's not simply about "oh I'm protecting against an adversary". The first basic thing these tools are going to give us is a mechanism to collaborate for a common goal and help prevent vulnerabilities from devious parties. Let’s acknowledge that there is data that needs to be kept private, not for any bad or ominous reasons. Certain data such as health records, financial records, must be kept private, but you still want to make use of it, and secure multi party computation is something that helps you to collaborate on private data.
Zero knowledge, on the other hand, is more individualistic in the sense that it protects the privacy of a single individual as opposed to a group of individuals. When it comes to multi party computation, zero knowledge helps for a single entity to communicate about something that they want to keep private. Let’s say I have assets, in traditional markets or in the cryptocurrency market. I can communicate what compliance these assets satisfy without ever revealing anything about the data. Zero knowledge let’s a single person demonstrate something about their private information. While secure multiparty let’s a group of partners with their own private data combine it in a useful way. Secure multiparty computation and zero knowledge are more like privacy enhancing tools.
There are a few obstacles. Based on our interactions with the industry, we have seen that people need to be educated on MPC. Next, they need to gain trust in the solutions. There is a need for a key application that will captivate the community (as blockchains were for zero-knowledge) and this will come from an application where privacy is crucial, there are no trusted intermediaries and MPC is the only solution.
Carmit: MPC is an enabler since it can allow and enable sorts of communications that are forbidden, sometimes by law. Think, for instance, of hospitals that would like to collaborate. A hospital wants to run some joint research on patient records, but these records can’t be shared because these are protected under privacy laws. Nevertheless, using the tools of secure computation, they could run the protocol that will provide some statistics out of this data. These statistics can be extracted from the data and help to better understand, say, the causes for some diseases, or connection between different specific genes, for example. There are countless numbers of opportunities to explore using this tool.
SB: Thanks, Carmit, I think that's a good example, and makes it very relatable for people. I suppose there are people reading this who might be finding out about secure multiparty computation for the first time. I would imagine readers also want to know when we’ll be able to experience this technology, or when will we see any exposure to this technology?
Muthu: Multi-party computation and zero knowledge, the basics, the feasibility, were established in the 80s, close to four decades back. But looking at the state of the art today, there has been more significant progress in zero knowledge within the last half decade, in just implementing zero knowledge proofs. The crucial element here has been blockchains and cryptocurrencies. Zero knowledge “naturally” fits into what you want to get out of blockchains. You do want to commit on blockchains and, often, you want to keep them private, which means they are in some encrypted form on the blockchain.
The simplest example: I want to prove I have X dollars in a cryptocurrency, but I don't want to share that information, I want to just prove that I have it. We call this proof of asset or proof of reserves. The tool that you want here is zero knowledge proofs. This has gotten the interest of the industry and has reinvigorated the interest in the academic world and now there's tremendous progress in the implementation of zero knowledge proofs. So, I think blockchains have been a good catalyst for zero knowledge proofs, and multi-party computation has also benefited from this.
Carmit: The blockchain community has indeed created many opportunities, and many new applications that didn’t exist before. Zero knowledge was the tool that was really pushed by this technology and this community. It was pushed because people want to protect the privacy of their data. They want to protect the privacy of the transactions.
As for MPC, I think one way to push people or companies to use it is by privacy regulations. It is very important that the governments, and the regulators, in different countries will be aware of the importance of using these tools, and will use the law, to provide or to regulate this so that people will be, I wouldn't say forced, but will be maybe encouraged to use these tools. GDPR regulation is an example, in the European Union, and there are other examples, but we still have a long way to go towards regulating the many uses of databases and data that are out there. And once it happens it will naturally create an incentive for companies to use these tools.
SB: It's interesting, and very clear and obvious to see the immediate benefits, and the potential for enhancing privacy over people's data. Carmit, Muthu, thank you very much for taking the time to talk to us.