Xinhua Full Media+丨 From "Algorithm Selection" to "Selection Algorithm": my country's network space governance has entered a new stage

Xinhua News Agency, Tianjin, August 30th (Reporter Liang) The algorithm is deeply embedded in people's lives, but the negative impact of big data killing and personal privacy leakage has caused people to worry about algorithm abuse.

On March 1 this year, the "Recommended Management Regulations on the Recommendation of Internet Information Service Algorithms" was officially implemented. For half a year of the new regulations, what new progress has been in the field of algorithm governance? In 2022, the guests and other experts and scholars gave their answers by the guests and other experts and scholars.

How does the algorithm governance have been effective for half a year?

As the basis of the digital economy, algorithms have been widely used for many years. Regarding the chaos of algorithms that appeared in recent years, Liu Xianglong, deputy director of the National Key Laboratory of the Software Development Environment of Beijing University of Aeronautics and Astronautics, believes that this is caused by artificial rules to data -driven. "The current algorithm and big data are coupled. Due to the uncertainty of the data and the boundaries of the algorithm capacity are unclear, the algorithm behavior will be uncontrollable."

Starting on March 1, the "Recommended and Management Regulations on the Recommendation of Internet Information Service Algorithms" (later referred to as "Regulations") jointly issued by the National Network Information Office and other four departments was officially implemented. This is my country's first Internet information service field with targeted algorithm recommendation rules and regulations. Several interviewees said that the "Regulations" introduced the gaps in the field of Internet governance, and my country entered the stage of systemic governance in the algorithm.

According to Liu Xiaochun, the executive director of the Internet Rule of Law Research Center of the Chinese Academy of Social Sciences, the "Regulations" landing means that my country is at the forefront of the world in algorithm governance exploration. "What is more valuable is that the" Regulations "respond to practical problems such as algorithm content governance and protection of minors. It is a governance plan with Chinese characteristics."

At the forum, many guests shared the results of the "Regulations" for half a year: Carry out the special operation of the 2022 "Clear Range · Algorithm Comprehensive Management" to conduct on -site inspections on some Internet companies; release the name and filing number of the first batch of domestic Internet information service algorithms and filing numbers Including a number of large Internet companies and products; clearly proposed that "if the user chooses to close the algorithm recommendation service, the algorithm recommendation service provider should immediately stop providing related services."

In addition to actual governance operations, Liu Xianglong believes that the guiding impact of the "Regulations" is equally important. "This releases a signal of the top -down system governance of the algorithm. Internet companies should actively regulate their behavior and ensure that users' rights and use is fair."

On August 28, at the National Convention and Exhibition Center (Tianjin), the participants visited the exhibition at the China Network Civilization Conference in 2022. Xinhua News Agency reporter Sun Fanyue Photo

How to resolve the algorithm "Black Box" technical problem?

At the algorithm governance sub -forum, Shen Huawei, director of the Data Intelligence System Research Center of the Institute of Computing Technology of the Chinese Academy of Sciences, pointed out that people have some misunderstandings about the filing of Internet information service algorithms that are implemented in synchronized with the Regulations, mainly the content of the filing. "The contents of my country's algorithm filing include three parts: the main information, algorithm information, product and functional information of the algorithm, which does not involve the data and code used by the algorithm."

Several experts and scholars interviewed believe that the opaque data, decision -making mechanisms and design intentions have made algorithms that rely on data for deep learning and analysis. Taking a taxi as an example, each passenger hopes to get the car as soon as possible, but the platform's single algorithm may consider how to distribute the vehicle to achieve the best global capacity, and then the subsidy design can achieve the largest profit. The algorithm optimization goals and user demands are between the algorithm optimization goals and user demands. There are certain deviations.

As a result, the algorithm recommendation service provider, as an important part of my country's algorithm governance system, is of great significance to the supervision of the algorithm before and in the event.

According to Shen Huawei, algorithm recommendation service providers should conduct algorithm security self -assessment before the algorithm launch application, eliminating possible risks. At present, before the technological development and supervision tools have not yet matured, the effective supervision of the "black box" algorithm has a way to go. The algorithm safety assessment can be used as an important supplement to algorithm governance.

On August 28, at the National Convention and Exhibition Center (Tianjin), the participants filmed the exhibition at the site of the China Network Civilization Conference in 2022. Xinhua News Agency reporter Li Ran

How does the algorithm get along with people in harmony?

At the forum, a number of experts said that the guidance and regulations on the governance of algorithms released in my country are actually "behavioral specifications" instead of technical underlying levels used for algorithms.

At the moment when the problem algorithm is frequent, is it in the process of getting along with the algorithm to be passive? Can users find initiative in the practice of algorithm governance?

Liu Xiaochun pointed out that the "Regulations" clearly required that "the user chooses to close the algorithm recommendation service, the algorithm recommendation service provider should immediately stop providing related services." "This has improved the active position of users in the algorithm service, which means that users can have the right to be selected, guided, or even threatened by algorithms."

Liu Xianglong believes that people and algorithms are essentially a relationship between "commissioning" and "agent" -the users hand over personal data to the algorithm and entrust their agents to make their best decisions. But in fact, due to the technical threshold of the algorithm, users lack corresponding technical identification ability. "This requires the main responsibility of the algorithm service provider of the‘ middle business ’, practicing social value, making the algorithm data more transparent and open, and making users trust.”

In addition, in the context of the systemic governance of the algorithm, users as an important participant must also rationally cognize algorithms, neither fear of technology nor relaxing the reasonable requirements of the algorithm, form a joint force with the regulatory authorities, and promote from the "algorithm selection" toThe transformation of "Select Algorithm".[Editor in charge: Jiao Peng]

- END -