From internet slaves to internet citizens, how will privacy computing change the way the internet operates?

This article was published on PlatON, author: LatticeX Foundation.

Since the concept of Privacy-preserving Computation was proposed in 2016, after five years of development, driven by relevant laws and regulations on data privacy protection in various countries, coupled with the increasing digitalization of human society due to the global pandemic, data has increasingly become the primary energy driving economic development. This has made privacy computing gradually the most fundamental and important underlying technology in information technology.

Privacy computing is a computable model and axiomatic system for privacy measurement, privacy leakage cost, privacy protection, and the complexity of privacy analysis when the ownership, management, and usage rights of private information are separated. Privacy computing allows data or computation methods to collaborate while remaining in an "encrypted" state, without disclosing to other cooperating parties.

This method of utilizing data in an encrypted state to realize its value is currently and will fundamentally change the operation of the internet and human society.

From Plaintext Data to Ciphertext Data

As a basic component of information networks, data faithfully records all information and trajectories of human production and life in a digital state. With the widespread adoption of emerging technologies such as cloud computing, big data, and artificial intelligence, a large amount of value has been mined from data, continuously improving and even transforming human life.

Relying on data to generate value is the basic business model of internet giants, and its premise is the collection of user information. User information data essentially constitutes an important part of individual rights, and the privacy rights of personal data have increasingly gained attention and are protected by law in various countries.

Therefore, the long-concerned issue of data security has gradually shifted from traditional fields such as hacker prevention, disaster recovery, and data restoration to focusing on how to protect user privacy while mining data value.

Currently, in almost all application scenarios, data is shared and used in plaintext form. Since data exists in a digital form rather than a physical one, lacking exclusive characteristics, the "what you see is what you get" nature makes it practically impossible to distinguish between the "use" and "ownership" of data.

This form of storing data in plaintext has, in fact, created a natural opposition and contradiction between data sharing and privacy protection. Even the model of releasing data after removing sensitive information can still identify user identities through quasi-identifiers.

The existence of this contradiction is a major development obstacle for industries with a high demand for data applications. For example, in the AI industry, on one hand, massive amounts of data are needed to feed and train AI to become smarter and more powerful; on the other hand, strict requirements for data privacy protection severely limit the amount of data that can be used compliantly.

The proposal of privacy computing provides a win-win solution to this contradiction.

Through privacy computing, data can circulate and be shared in an encrypted state for computation, completely separating the states of "use" and "ownership." Data becomes "invisible but usable," allowing for complete protection of data privacy while still enabling the realization of data value.

Thus, it can be foreseen that in the future, as privacy computing technology becomes the underlying information technology, data residing in every corner of society will be shared and used in a completely unreadable ciphertext state unless authorized by the data owner, fundamentally solving the issue of data privacy protection.

From Information Network to Computing Network

Once data transitions from "plaintext" to "ciphertext," the internet as we know it will leap from an information network based on data exchange to a computing network based on computational interoperability.

Today, internet giants possess vast amounts of user data and continuously extract value from it through data mining to achieve commercial revenue. This state will be thoroughly disrupted in the future.

With the support of privacy computing technology, new collaborative paradigms will emerge.

The traditional internet can be defined as an information network oriented towards direct processing of data. Local data is collected by various applications (often silently and without the owner's permission) and transmitted to the cloud, where it is processed and mined to generate commercial value. However, data owners have not authorized this process and have not received corresponding returns from the commercial value created.

A more serious situation is that these plaintext data are directly sold after being collected, leading to significant privacy data leaks and causing severe social impacts and economic losses.

Privacy computing will allow users to keep their data locally and participate in computations in an encrypted state, uploading the computation results to the cloud for the computation initiator. Throughout the process, plaintext data is never disclosed, fundamentally protecting data privacy. At the same time, the value created by the data can be measured through reasonable mechanisms, allowing data owners to receive fair returns.

The ultimate direction of the evolution of the internet is to transition from an information network form that transmits data from endpoints to the cloud, towards a computing network oriented towards data sharing and collaborative computation.

As advocated by the distributed privacy AI network PlatON, "everything is computable."

From Single-party Computation to Multi-party Computation

As the object of computation, "data" possesses very unique properties that differ from traditional production factors. In addition to being "what you see is what you get" in plaintext, it also has "inherent" partial public attributes.

Specifically, data from a single source is actually insufficient in credibility and value; data that can truly create sustained value should feature multi-source, multi-dimensional, and multi-party continuous operation.

Therefore, in a digital society where data is the basic production factor, analysis and computation oriented towards data will not be limited to a single source of supply.

For example, the AI for autonomous driving must be "fed" with data covering various scenarios, road conditions, and rules from multiple sources. It can be simply understood that an AI trained with driving data primarily from open roads in Area A will not be suitable for autonomous driving in Area B, which mainly has complex road conditions.

If the data owned by "Area A" and "Area B" also has issues of data ownership control, then either the autonomous driving AI trained with single-source data will be fundamentally unusable due to the limitations of the data supply, or a way must be found to jointly train the data from "Area A" and "Area B" without disclosing the data.

With the core essence of data possessing public attributes, the value of data must be realized through collaboration among multiple parties. Therefore, computation oriented towards data cannot be achieved as before through centralized single-party computation; instead, it should output computation results through multi-party computation in an encrypted state without leaving local data.

From Network Slaves to Network Citizens

In real society, citizens, as the basic units of the state, possess fundamental rights protected by national laws and guaranteed by the government. These rights include the right to participate in public social life as stipulated by law. For instance, civil rights include property rights, personal rights, privacy rights, and so on.

The relinquishment of certain rights by citizens constructs the power of the state.

However, in the online world, users utilize public digital products provided by network service providers (such as internet giants) and completely provide their personal data to these digital products involuntarily. Network service providers have complete and absolute disposal rights over user data.

For example, they can delete a user's account without permission.

They can mine user data to extract commercial benefits without paying any returns to the user.

Thus, fundamentally speaking, network service providers wield state-level power similar to that of real-world governments, but this power is acquired coercively, without the premise of relinquishing rights by service users.

In other words, in the online world, each of us users has no rights at all. Regardless of our identities in the real world, as long as we connect to the internet and use network services, we become slaves to these network services—contributing value, relinquishing rights, without any returns.

Only when privacy computing truly becomes the foundational technology of the internet, and all network service providers adhere to the privacy computing paradigm to provide services, allowing our data to be stored in ciphertext, separating "ownership" and "use," and enabling us to participate in computations entirely according to our will, can participants in the online world truly become network citizens with fundamental rights.

As a newly emerging information technology, privacy computing will not simply bring new applications; it will fundamentally change the basic operational model of the internet and allow network participants to use various network services with a new identity, improving living standards, enhancing work efficiency, safeguarding data rights, and constructing the online world.

The LatticeX Foundation, headquartered in Singapore, aims to build a fully decentralized computing interoperability network that returns data sovereignty to users, protects data privacy, and realizes data value exchange. It seeks to promote the trading of data usage rights while protecting data sovereignty and privacy, funding various academic research, managing applications, and providing financial support for quality projects. The LatticeX Foundation is a major supporter and promoter of the distributed privacy AI network PlatON and the financial network Alaya.