Matrix Element Sun Lilin: How does privacy computing solve the problems of data ownership and transaction?

This article was published in Jiemian News, guest: Sun Lilin, written by Si Linwei.

The COVID-19 pandemic has caused a significant impact on society, with technological tools permeating every corner of life, blurring the lines between public and private spaces. Social issues related to personal privacy are emerging one after another, such as requiring facial recognition to enter residential areas, real estate companies categorizing clients using facial recognition, and certain apps knowing users' interests better than the users themselves… The omnipresent cameras raise the dilemma of whether they make people safer or sacrifice privacy.

While surrendering personal information to maintain public safety, how can privacy and personal data be protected from misuse? How can the security risks and issues of commercial benefit distribution arising from this be resolved?

Professor Shi Zhan from the Diplomatic Academy mentioned a term in his book "Breaking the Cocoon"—privacy computing. "All data should reside on one's own terminal; personal data entering an algorithm is somewhat akin to entering a black box, and technically it can ensure that all computing processes of the data are encrypted, with no one knowing what the data actually is; it can only provide a computational result." Shi Zhan believes that "once I commit to allowing you to use my privacy in a certain way, I should receive dividends." He thinks that the definitions of personal rights, property rights, and so on that we were familiar with in the past may be iterated, and various new rules will emerge.

What exactly is privacy computing? How should the rights and responsibilities among individuals, companies, and governments regarding big data composed of personal information be distributed? Jiemian News interviewed Sun Lilin, founder and CEO of Matrix Elements, on this topic. Sun Lilin is a founding member of Chinaledger and vice chairman of the technical committee, as well as a member of the Financial Technology Committee of the China Securities Investment Fund Association. He has long been engaged in payment and clearing financial infrastructure work, with in-depth understanding and rich experience in how blockchain transforms the traditional financial industry and information transmission.

Founded in 2014, Matrix Elements positions itself as a provider of privacy computing and blockchain technology, aiming to solve data security and privacy protection issues in the collaborative computing process of the digital age. In 2016, Matrix Elements received an investment of 150 million yuan from Wanxiang Holdings and has in-depth cooperation with research institutions such as Shanghai Jiao Tong University, Fudan University, Wuhan University, and the Chinese Academy of Sciences. Just last month, the Shanghai Municipal Science and Technology Commission announced that Matrix Elements' wholly-owned subsidiary, Shanghai Zhenfang Technology Co., Ltd., was approved as a "Key Laboratory for Data Privacy and Security Computing Enterprises in Shanghai" and "Shanghai Distributed Privacy Artificial Intelligence Technology Innovation Center."

In Sun Lilin's view, data itself is a result computed collaboratively by multiple parties, and the ownership of data is indeed difficult to define. The greatest value of blockchain is that it solves the fundamental clearing and settlement issues, transforming multi-level data custody into a distributed and highly efficient system. As a factor market, data must not be controlled by any single institution or platform. The market for privacy computing will likely see a model similar to government-authorized licenses in the future.

Below is the interview record, with some edits.

Sun Lilin, CEO of Matrix Elements

1. How to Ensure That Technology Holders Do Not Misuse Our Privacy Information

Reporter: Can you explain in simpler terms what data is?

Sun Lilin: Ackoff (1989) proposed a famous model—the DIKW model, where D stands for Data, I for Information, K for Knowledge, and W for Wisdom. In this model, the lowest level is data, followed by information, then knowledge, and at the top is wisdom, which reflects the original way humans understand things.

Data is actually the interaction process between humans and the external world, from which humans abstract information, then knowledge, and finally elevate it to wisdom.

Data is manifested in services or generated during the service process, serving as a carrier throughout the entire service process. Simply put, objective observations and recordings of everything and behaviors constitute data, so humans are independent of data.

Reporter: Many residential areas now require facial recognition for entry, leading many people to store their facial information in these areas. Generally, when entering office buildings, people only need to register personal information, including ID numbers and phone numbers. Do you think this method is reasonable?

Sun Lilin: This indeed involves a good example of privacy. For instance, registering with an ID to enter an office building reflects the relationship between visitors and property managers. As a private territory, the office building certainly has the right and motivation to protect its space by prohibiting unauthorized personnel from entering. This right is legitimate. However, the question is, to whom does this right belong: the property owner, the property management, or the renting company?

But the subsequent question is how to ensure the security of information when they share it.

Reporter: After I provide personal information at the front desk, does the corresponding company have the capability or obligation to protect my privacy?

Sun Lilin: It does not necessarily have the capability, nor do I think it considers itself obligated to do so.

The tacit agreement now is that the property management has the authority to decide who can enter the space, rather than the renting company. Because while the property management has the power, it also has corresponding obligations and responsibilities to bear the risks.

However, the vast majority of properties do not have the capability to truly do this, nor do they have sufficient motivation to protect visitors' information security, as their primary motivation is to protect their private space.

When we enter an office building and register information, we establish a contractual relationship with the property, which serves both as a deterrent and a means of statistics; if something happens, the property can find the corresponding personnel.

But the common situation now is that the property does not make another commitment: I will not use your information for other purposes.

Thus, information leakage inevitably occurs repeatedly.

Reporter: Can technology, especially blockchain, improve the issue of information leakage?

Sun Lilin: Technological means can solve this problem to a considerable extent, replacing or becoming the infrastructure for public governance in human society.

However, when technology is fully integrated as a new paradigm of public infrastructure, it not only solves some problems but also brings deeper issues: who controls this technology.

In fact, we have completely disclosed our rights to the so-called technology holders. But how can we ensure that these technology holders do not completely misuse our privacy information? How can we ensure that large platforms like Alibaba and Tencent do not misuse our data?

2. Data is a Result Computed by Multiple Parties, and Its Ownership is Difficult to Define

Reporter: What does privacy computing really mean? Can it truly protect privacy?

Sun Lilin: Regarding privacy computing, the term I initially used when founding Matrix Elements was collaborative computing, which essentially means human collaboration.

The fundamental paradigm shift in data computation is even more profound than that in finance. Data computation shifts a large amount of computation and transaction processing from centralized clouds to the edge, to individuals and institutions, giving us more capability to handle computation while significantly lowering the threshold.

For example, the privacy computing product Rosetta has the slogan of enabling people who do not understand cryptography to use cryptographic technology. What we do with AI products is to allow those who do not understand AI to use it. Therefore, what Matrix Elements truly aims to do is to enable those without data to use data.

Privacy computing solutions are particularly suitable for the medical field. I used to be very radical in believing that case information belongs to the patient and not to the hospital, but later I realized this was wrong. Why? The patient is the subject of the illness; without the knowledge of the doctor applied to my body, the patient would not know what illness they have.

Patients cannot write case notes; they only know they have a cold and a runny nose because they have some understanding of colds. However, if there is a more serious illness, the patient would not know; only through the doctor and equipment can it be detected.

Thus, the vast majority of data today, in a significant sense, is the result of computation, and this computation incurs costs from others, including equipment, labor, etc.

Therefore, data itself is a result computed collaboratively by multiple parties, and the ownership of data is indeed difficult to define.

Reporter: Can blockchain play a role in defining data ownership?

Sun Lilin: Data has a public nature, and data computation is essentially multi-party computation, which also applies to privacy computing.

The greatest value of blockchain is that it solves the fundamental clearing and settlement issues, transforming isolated multi-level data custody into a consistent, distributed, and highly efficient input.

Reporter: Some experts point out that the essence of blockchain is a public ledger, and is there a contradiction between its public nature and privacy protection?

Sun Lilin: Privacy computing and blockchain are two-dimensional products that complement each other as tools.

Blockchain itself is not encrypted; Bitcoin only uses hash signatures, and its level of encryption is even lower than that of some high-end mobile phones. The connection between blockchain and privacy computing is that it merely provides clearing capabilities for privacy computing. The privacy processing part is mostly done locally.

3. The Game of Interests Between Institutions Constitutes the Business Model

Reporter: If blockchain is a public infrastructure, what role can companies like Matrix Elements play in privacy computing?

Sun Lilin: There is a particularly classic saying: unawakened rights are not rights. When I am not aware that this is my right, I do not care whether this right is taken away. Because privacy rights generate benefits after they have utility, people awaken in the face of benefits and will protect their privacy rights.

Once people awaken in the face of benefits, this matter becomes very critical. Therefore, when designing the business model for Matrix Elements, I believe that 2C (consumer-facing) privacy protection will only constitute moral pressure for a considerable time, rather than a business model. Only the information protection between institutions constitutes a business model.

The game of interests between institutions constitutes the business model. Institutions will first realize the value of this asset, and they have sufficient scale. The benefits that individuals can monetize are too small; individuals do not constitute an entity for equal games, so there is no model.

Reporter: So, is the 2B business unrelated to or mutually promoting the 2C business?

Sun Lilin: 2C is a catalyst. The development of 2C business in China has a default premise: its material supply is unlimited, such as shared bicycles, which can be produced infinitely; or Meituan delivery, where restaurants can supply infinitely. The larger the restaurant, the more diners there are, and the cheaper the supply chain becomes. Under the premise of unlimited supply, scale can be expanded.

In today's blockchain industry, the vast majority of people only regard it as a 2C application and do not treat it as infrastructure.

The situation in the infrastructure field is the opposite of that in the C-end; all infrastructure businesses have limited supply. For example, the more electricity you use, the more expensive your electricity bill becomes; the more cars on the highway, the higher the tolls. The same goes for blockchain; Ethereum is a typical example. The more people use it, the higher the gas fees.

The reason for this situation is the natural limitations that infrastructure faces; the public infrastructure for all humanity can only provide so many services.

Returning to the financial field, the supply of clearing capacity at the clearing layer is limited, not as unlimited as everyone imagines, which is why many consumers' shopping carts cannot be cleared in time during Double Eleven. Even if many hardware servers are stacked, databases optimized, and TPS (transactions per second) improved, the clearing capacity of UnionPay's clearing library is limited.

As an infrastructure provider, UnionPay's improvement in payment performance is stage-limited, while the iteration and update of applications far exceed that of infrastructure, and the cycle of infrastructure evolution is much longer than that of applications.

And infrastructure means the need for regulation. Before the massive rise of e-commerce, the internet only had the HTTP protocol; when e-commerce payments began to become popular, only then did HTTPS and SSL protocols emerge. Only when SSL and HTTPS are regulated can humanity trust them. The essence of all human transactions is trust.

Today, from a global perspective, if we recognize data as a factor market, then the factor market must not be controlled by any single institution or platform, regardless of which tech giant it is.

Therefore, I believe the business model of privacy computing has also returned to the era of building telecom operators in China over 20 years ago, where there will definitely be operators authorized by the government to provide public infrastructure services.

Everyone can claim rights over the data they can control and process. However, if we attempt to establish a network for exchange and sharing, we cannot directly obtain the data and rights of other participants; all parties participating in the network must cede part of their ownership and usage rights over the data to a third-party operator for the network to function.

For Matrix Elements, we do not seek the status of an operator; we are committed to doing technology well and providing technical solutions for future large data operators.

Reporter: What work have you done so far?

Sun Lilin: We started working on privacy computing in early 2017, initially believing that this direction was definitely correct, but the path was unclear. It wasn't until last year when new infrastructure began to gain popularity that I felt the path was basically clear. Currently, we have the basic technical capabilities to meet some fundamental needs and provide the infrastructure for information exchange between enterprises.

Matrix Elements RosettaFlow Privacy AI Platform

Reporter: You previously worked at UnionPay and entered the blockchain industry in 2013. Is the relationship between finance and blockchain one of mutual promotion or replacement?

Sun Lilin: Blockchain is inherently financial infrastructure. What Nakamoto expressed in the Bitcoin paper is also this meaning; he wrote in the abstract that to completely solve the issues of transaction traceability and fraud in the era of online transactions, the first thing to address is the most fundamental clearing mechanism of financial infrastructure.

The essence of blockchain is merely a middleware for services; it is transaction-oriented. Blockchain is actually just a hand and foot for humanity, helping humans work.

Matrix Elements, engaged in distributed AI privacy computing, is not transaction-oriented but data-oriented. The goal of Matrix Elements is not to replicate hands and feet but to replicate the brain. AI has three levels: perception—cognition—decision-making. Today's blockchain is still at the stage of perception, and what Matrix Elements aims to do is to use blockchain means to gather massive data to assist human cognition and decision-making, ultimately becoming an "external brain."