■ China Economic Times reporter Zhang Na

On March 25, Jing Xiandong, Chairman and CEO of Ant Group, attended the 2024 Annual Meeting of the China Development Forum. When he delivered a speech at the symposium on "Digitally Empowering Industrial Transformation", he said that the application of data element circulation has brought Industrial value. At present, the circulation of data elements is moving towards the third stage, that is, to achieve credible circulation between industries and regions. However, security, cost and computing complexity restrict the further development of the value of data elements. Ant Group is exploring the next generation of private computing technology, striving to be as efficient and easy to use as plaintext computing, allowing the flow of data value to be available like tap water, achieving security, convenience and low cost, and grasping the exciting era of data elements.

蚂蚁集团董事长兼首席执行官井贤栋

In recent years, the in-depth development of China's digital economy has promoted exponential growth in data scale. At the same time, the isolation of data around scenarios has also become more prominent. In Jing Xiandong's view, the arrival of the era of General Artificial Intelligence (AGI) requires solving the problem of large-scale data circulation under data isolation. The trustworthy flow of data elements requires innovative technological support, including the Internet of Things, blockchain, private computing and artificial intelligence.

At present, data element technology has been implemented in the industry and created new value. For example, he said that the power battery industry has a long chain and faces problems of difficulty in echelon recycling, difficulty in valuation and pricing, and difficulty in industry collaboration. Using blockchain and IOT technology, power batteries can be implanted into a trusted chain module,"connecting the chain with power", achieving transparency and credibility of operating data throughout the life cycle, and combining privacy computing technology to ensure data flow between upstream and downstream enterprises and Financing, insurance and other institutions are safely transferred to form a credible industrial collaboration network.

The circulation and utilization of data elements have been applied in points and are entering a stage of rapid development. He took the construction of a water supply network as an example. The first stage of the circulation of data elements is a data island, just like their own wells. Enterprises have the ability to self-research data and produce and sell it; the second stage is to realize point-to-point circulation of data elements, just like bottled water., lack of infrastructure support; in the third stage, data elements are credibly transferred in industries and regions, just like urban water networks; The last stage is the trusted transfer of data over a wide area, like a comprehensive water conservancy project, which achieves trusted transfer and interconnection across industries, regions and clouds, truly benefiting thousands of businesses.

Jing Xiandong believes that at present, the circulation of data elements is moving towards the third stage, that is, to achieve credible circulation between industries and regions. However, security, cost and computing complexity restrict the further development of the value of data elements.

For example, in the data trading market, the risks brought by plaintext circulation are high, data leakage and abuse lead to rampant illegal transactions, and sellers are "unwilling to circulate" and "dare not circulate"; in the buyer's market, ciphertext data exchange is inefficient, and circulation costs may exceed the value of the data itself, and the buyer is "unwilling to pay". In addition, the rapid development of artificial intelligence technology has also put forward higher requirements for private computing. Due to cost constraints, private computing is still in the laboratory stage in high-level artificial intelligence such as large model applications.

Jing Xiandong said that Ant Group is committed to exploring the next generation of private computing technology and striving to be as efficient and easy to use as plaintext computing. Among them, one exploration is to launch trusted secret computing to achieve improvements in dimensions such as low cost, high security, high performance and high reliability.

In terms of cost, at the same high security level, trusted secret computing will control the cost within 10 times the cost of plaintext distributed computing.(In comparison, multi-party security computing is 1000 - 100000 times), striving to make the performance and cost of private computing approach plaintext computing; in terms of security, due to the dual security guarantees of cryptography + trusted hardware, it can reach a high level of attack and defense verification. Security level; In terms of performance, the prediction of the million-level parameter CNN model can be completed in the sub-second level, and the SQL analysis of hundreds of millions of pieces of data can be completed in 10 minutes.

"We hope to make private computing services universally inclusive and become a catalyst for the trustworthy flow of corporate data elements." Jing Xiandong proposed that to promote the implementation of more advanced privacy computing technology industries, we must establish a data technology development and application ecosystem.:Technology-based digital vendors continue to explore and innovate technology and lower technical thresholds; application-based digital vendors purchase secret sky computing services like cloud services; small and micro enterprises participate in data circulation through secret SaaS services.

According to reports, Ant Group continues to focus on private computing technology while insisting on open sharing. At present, Ant Group's core technologies have all been open source, including the "lingo" trusted privacy computing framework, the trusted execution environment operating system Occlum, etc. In addition, it has participated in the formulation of more than 60 domestic and foreign standards, including 9 international standards. Ant Group also works closely with more than 50 domestic and foreign universities and has established a "Private Computing Special Research Fund".

"Our vision is to make the flow of data value available like tap water, achieve safety, convenience and low cost, and grasp the exciting era of data elements." Jing Xiandong said. (This article is compiled based on on-site recordings and has not been reviewed by me)