Not long ago, Qiji founder & CEO Dr. Lu Qi was asked in a question-and-answer column about his current views on AI, and he said, “My enthusiasm and concern for AI mainly comes from the prospects it can bring to our society. .” In Lu Qi’s view, the core of AI is a “general ability to acquire knowledge and use knowledge to achieve goals”. And this is the most versatile and most functional general ability ever invented by human beings. Because knowledge is power, a power that can be applied to anything we want to do.
Looking back at the past few years, my country’s AI has gone out of the laboratory driven by data and landed in many fields such as finance and security. There is even a saying in the industry that “whoever gets data gets AI”. In 2020, the State Council listed data as the fifth element after land, labor, capital and technology to encourage the circulation and value expression of data. But unexpectedly, in the short term, AI companies have fewer channels to obtain data. On the one hand, after data has become a factor of production, individuals and the government have improved their awareness of the value and protection of data, and enterprises will more selectively open up and share data types and methods; on the other hand, data is one of the factors driving enterprise development. First, under the consideration of legal restrictions and self-interest, it is strictly controlled by various enterprises.
The further development of AI companies requires more data, but data cannot flow as smoothly as in the past, and AI applications are falling into the development bottleneck of data fragmentation.
Regarding how to find a balance between AI development and privacy protection, Zhang cymbal, a professor at the Department of Computer Science of Tsinghua University, gave two ideas: on the one hand, how to prevent privacy from being misused and abused. The second aspect is how to use technical means to protect the privacy of individuals or groups, including data security and so on.
The former belongs to the governance of artificial intelligence, while the latter is a technical problem.
At the same time, a group of people found that privacy and security computing with the characteristics of “data availability and invisible” may help AI companies get out of the data dilemma and open the door to data circulation.
What is Privacy Secure Computing
Privacy and security computing is a collection of technologies that ensure that data providers do not leak data to the outside world and cannot be maliciously attacked or obtained by other unauthorized persons during the process of data processing, analysis and computing, and can realize the safe circulation and utilization of data.
A classic question: Two millionaires meet on the street. They both want to know who is the richest in each other, but they don’t want to get to the bottom of each other. How can they know who is the richest without the help of a third party?
This is the “millionaire” question posed by Yao Qizhi, winner of the 2000 Turing Award in 1982. The problem raised by Mr. Yao and the solution he proposed have become a general direction in the field of cryptographic security, promoting the development and application of privacy-safe computing technology.
In the past two years, privacy and security computing has become a new industry that investors are optimistic about. According to the statistics of Prospective Economist, the number of newly established enterprises in my country’s privacy and security computing industry in 2020 will be 71, a year-on-year increase of 33.96%.
In the 12 months from May 2021 to the present, 8 financings of 8 companies in this track have accumulated over 1 billion yuan, and the average single round of financing amounted to over 100 million yuan.
It is worth noting that most of the company’s financing events occurred in 2020-2021, which also reflects that more and more investors are discovering the value of privacy-safe computing.
The rapid development of privacy and security computing is inseparable from the progress of algorithms and the substantial improvement of computer performance on the one hand, and it is also related to policies on the other hand.
Algorithms for privacy computing have made great progress in the past decade, including breakthroughs in areas such as differential privacy, federated learning, homomorphic encryption, and zero-knowledge proofs. The demand for computing power and communication bandwidth of private computing technology has also been greatly improved due to the development of computer systems and hardware. The technology of private and secure computing can finally begin to solve practical tasks, not just pure theoretical problems in the computer field.
In terms of policy, with the successive entry into force of three laws, namely the Cybersecurity Law, the Data Security Law and the Personal Information Protection Law, companies are forced to attach importance to and enhance data protection in the entire process of data collection, processing, use, and circulation. Hence profit.
The “14th Five-Year” Digital Economy Development Plan issued by the State Council in January this year clearly stated: “Encourage key industries to innovate data development and utilization models, and mobilize industry associations and research institutes on the premise of ensuring data security and user privacy. , enterprises and other parties participate in the development of data value.”
The issuance of this document may further accelerate the development and industry application of privacy-safe computing technology.
In the past few years, privacy and security computing has been extended from the medical industry to different fields such as finance and government affairs, and the entire industry has become more and more lively.
Ma Lan, vice president of Boiling Point Capital, shared with Leifeng.com her observations on privacy and security computing applications in recent years from the perspective of investors.
Malan noticed that in 2018, many financial institutions put compliance first, so many companies that took supervision as their entry point grew up at that time. After the government officially proposed to use data as a factor of production in 2019, data security has been elevated to an equally important position as compliance.
However, people found that there is a big data security problem when data is used for both assets and transactions. At this time, a group of people introduced privacy and security computing to help solve this problem.
Thanks to the emergence of a larger market demand, in 2020, the original enterprises in the privacy and security computing industry will exert their efforts, and at the same time, some new start-up companies will appear, and capital will also follow up. Therefore, from 2020 to 2021, privacy and security computing entrepreneurs will find new landing scenarios one after another, and even generate certain income.
In Malan’s view, although the privacy and security computing industry is in dynamic changes, it is developing towards a positive and more secure state as a whole.
The collision of AI and privacy-safe computing
Yang Qiang, executive director of AAAI International Advanced Artificial Intelligence Association, once told Leifeng.com that since 2019, he has clearly felt that artificial intelligence is difficult to implement, the application model is not universal, and the versatility of AI products is not enough. .
In recent years, many countries around the world have regarded data as a core asset. Data cannot be shared, forming data silos, further hindering the implementation of AI. He believes that data barriers exist in all walks of life. Only by breaking through the barriers and increasing the liquidity of data can the AI ecosystem develop better.
Under the requirements of regulations and policies, leading technology companies can obtain a large amount of data from multiple channels because of their mature products and huge number of users. Small and medium-sized enterprises do not have such conditions, and it is difficult to break through the data bottleneck.
Privacy and security computing is a way to break through industry data barriers. Privacy and security computing allows data to be secured in the process of cooperation, and data flow is naturally smoother.
At present, many subjects with a large amount of data must keep the data information strictly confidential, and at the same time cannot find a suitable processing method, the data is idle, and the value of the data cannot be exerted.
For example, a local government has detailed data of local residents and hopes to establish an intelligent infectious disease prevention and control system to prevent and control the epidemic. However, without technical support, it is difficult for the government to establish this system by itself. If external bidding companies help, there is a risk of leakage of residents’ personal data, and the government does not use the data in order to avoid the risk of data leakage, so the data cannot play its due role.
If a third party that provides privacy and secure computing services is introduced between the two parties, the data will not be directly circulated between the two parties, and the situation where the data owner is strong due to the possession of data in cooperation will be less likely, and the data flow will be relatively safer. .
Specifically, the privacy and security computing company will provide a corresponding platform, and the data provider will import the data authorization into the platform for model evaluation and optimization, and only output the value and calculation results of the data to the data demander after completion. During the whole process, the original data does not leave the privacy and security computing platform, and the data is only authorized for use within the platform.
In the process of cooperation between the two parties, data leakage can be avoided after the emergence of private and secure computing companies. However, how to ensure that privacy and security computing companies will not leak and abuse data?
Zhang Lintao, chief scientist of privacy and security computing company Yifang Jianshu, said that privacy and security computing is still a technology in the early stage of development, and there is still a lot of room for optimization in all aspects. However, in order to protect the privacy of data and information, there are many measures in the industry. .
Taking Yifang Jianshu as an example, the data optimized for training on its data platform is encrypted, and the key is owned by the data owner, and Yifang Jianshu cannot obtain the data; secondly, Yifang Jianshu’s multi-party secure computing and federated learning 、Trusted Execution Environment The three mainstream security computing methods have all passed the certification of the Institute of Information and Communications Technology, and the official endorsement proves their data security.
After seeing the value of privacy and security computing, many companies including Alibaba, WeBank, Ant Group, and Ping An Technology have actively deployed privacy and security computing and promoted technology applications. According to the survey data of the China Academy of Information and Communications Technology, about 44% of privacy and security computing products will enter the implementation stage in 2021, and the proportion will further increase; the proportion of privacy and security computing products in the research and development stage will decrease relatively, accounting for 19%.
In the foreseeable future, privacy and security computing may be deeply integrated with AI to help AI companies develop more rapidly.
Wing Fang Jian Shu is specially designed for AI’s problem-solving method
As Zhang Lintao said, there are still many problems in privacy and security computing technology to be solved.
First, privacy and security computing faces the problem of ecological barriers.
The technology between companies in the privacy and security computing industry is not interoperable. After the data model is output on one platform, it cannot be reused on another company’s platform, resulting in the emergence of new “data islands”.
Secondly, the willingness and market for data transactions are still immature, leading many companies to regard privacy and security computing as a cost item for security compliance. Only when scenarios are deeply integrated with privacy and security computing technology, business parties can turn cost items into revenue items after benefiting from privacy and security computing, and stimulate the willingness of business parties to participate sustainably.
In fact, in the past, many institutions made efforts to promote national data transactions, but due to technical limitations, the results were not ideal.
If combined with privacy and security computing, data transactions may be more efficient.
Leifeng.com has learned that Yifang Jianshu is planning to launch an “AI Taobao” based on privacy and security computing. According to Liu Shuo, its chief marketing officer, the platform can connect different AI demanders, suppliers, and data demanders with suppliers, enabling companies with different capabilities in the AI industry chain to play their respective strengths and meet different needs.
Specifically, the platform integrates domestic mainstream AI tools, and AI companies and data participants in the platform can access AI capabilities. The biggest difference from other platforms is that this platform protects all data, AI models of data sources.
“Yifang Jianshu is a data intelligence company with zero data. It does not own data. It only provides tools to manage data, and allows customers to process and process data under authorization to obtain data value.” Zhang Lintao introduced to Leifeng.com.
The reason why Yifang Jianshu has such a plan is related to its many years of deep cultivation in the industry and its long-term observation of the business development of companies in different fields.
Since its establishment in 2016, Yifang Jianshu has devoted itself to the development and application of privacy and security computing. At present, its business has expanded from medical care to government affairs, finance, marketing, science and other fields. In past cases, Yifang Jianshu has used privacy and security computing technology to solve practical problems in different scenarios:
Using privacy and secure computing technology, Yifang Jianshu helps companies with “drug-cell-gene” databases, such as Gewuzhihe, to reach supply and demand cooperation with AI pharmaceutical companies, biomedical R&D technology companies, such as Suikun Intelligent, and help data owners The user separates the right to use and ownership of the data, and can use it with confidence; for the bidding scenario, Yifang Jianshu has built an AI verification platform, which not only protects the data of the bidder, but also protects the AI enterprise of the bidder. Model. Not only is it applied to the bidding selection of AI demanders, but the AI verification platform can also be applied to technical competitions to achieve true “technical scoring” for AI.
Due to the complexity of implementation and delivery, the huge amount of computation, the low level of customer acceptance, and the need for full confidentiality, privacy and security computing technology has only just begun to be applied in more and more fields. With the advancement of technology, there are still many scenarios with opportunities for privacy and security computing to show its strengths.
Taking the automobile industry as an example, in the intelligent networked automobile industry that has emerged in recent years, many autonomous driving companies have emerged to provide assisted driving capabilities for car companies, such as Baidu Apollo and BYD, Momenta and SAIC. At the same time, some people questioned that self-driving companies may collect a large amount of user and road data through mass-produced cars in providing services for car companies, thus there is a risk of data privacy leakage. If privacy and security computing is introduced in the process of cooperation between the two parties, it may prevent the autonomous driving company from obtaining sensitive user information.
Summarize
Professor Song Xiaodong, known as the “Godmother of Computer Security”, has publicly stated that all computing in the future will be private computing.
With the acceleration of digital transformation and upgrading of various industries, the driving role of data in the development of the industry will become more and more obvious, and the flow of data will also be more restricted.
At present, many companies have proposed different technical routes to improve the degree of privacy and security computing security compliance. As privacy computing is gradually applied to more scenarios, to make up for current shortcomings, privacy and security computing may usher in a brighter future.
Leifeng Network Leifeng Network Leifeng Network
This article is reproduced from: https://www.leiphone.com/category/industrynews/70X6gqFxkS1kIP6J.html
This site is for inclusion only, and the copyright belongs to the original author.