Recommend the algorithm service to create a benign rule of law track

With the development of artificial intelligence technology, algorithms are increasingly integrated into all areas of digital economy and digital life. As the law of operation of the digital world, the norms and supervision of algorithms should be committed to the value orientation of the new order of digital civilization, and support the protection of the Chinese digital economy on the world stage. The Central Committee of the Communist Party of China issued the "Outline of the Implementation of the Construction of the Law Social Construction (2020-2025)", which proposed to formulate and improve the standardized management measures for new technology applications such as algorithm recommendation and in -depth falsification. ") In this context, it came into being. The "Provisions" is regulated from the design level recommended by the algorithm, covering hot issues such as algorithm discrimination, "big data killing", and induction of addiction, and proposes a Chinese solution for comprehensive governance of algorithms. The concept of transmission is clear: by advocating technology upward, realizing the traceability of algorithm behavior, the strength of algorithm supervision, and the effectiveness of the laws and regulations, thereby promoting the systemicization of algorithm governance and modernization of network space governance.

Adhering to the positive governance concept of technology

The "Regulations" clarify that the algorithm recommendation service provider should adhere to the mainstream value orientation, optimize the algorithm recommendation service mechanism, actively spread positive energy, and promote the application of algorithms to be good. This concept runs through: algorithm recommendation service providers shall not use algorithm recommendation services to engage in illegal activities or spread illegal information; no news information published by units published in units within the scope of non -state regulations shall not be generated; Public opinion, avoidance supervision and management, monopoly and unfair competition behavior. Algorithm recommendation services can provide more personalized and convenient information acquisition experience according to user interests, hobbies and needs. However, the lack of regulatory algorithm recommendations can also cause many issues such as algorithm black box, algorithm discrimination, algorithm indulgence, etc. How to calibrate the establishment of technological innovation and development within the legal framework requires us to explore a Hengping law that encourages innovation and protection rights.

The first "mobile phone traffic hijacking case" tried by the Hangzhou Internet Court fully considered the subjective intention of using the mobile phone app to wake up the strategy to implement the algorithm recommendation service behind the traffic hijacking. By effectively balance the diversified legal interests such as the right to know the right to know and the autonomous choice, and the value of market competitive rights and interests, the defendant finally ordered the defendant to stop the implementation of unfair competition and compensation for economic losses. This case makes a vivid instance interpretation and evidence for Article 4 of the "Provisions" that "providing algorithm recommendation services should respect social morality and ethics, and abide by business ethics and professional ethics." Provide more new ideas.

Pay attention to the combination of "rule of law" and "autonomy"

With the development of the Internet economy, online platforms have become an important subject of online operation through algorithm design, deployment and application. Therefore, the phenomenon of "algorithm power" formed by algorithms such as algorithm recommendation services must clarify the platform as an important "governance subject" and "governance object". In other words, the governance will be deepened to the underlying architecture of the platform, and the platform will be the "maker and executor of the governance rules" under the supervision of the rule of law.

The highlight of the "Regulations" is that on the basis of following the technical logic of the algorithm, the algorithm accountability mechanism in the platform economic scenario has been clearly constructed, and the "technical neutral" veil of algorithm automation decision -making has been pierced in legislation. For example, Article 8 of the Regulations: "Algorithm recommendation service providers shall regularly review, evaluate, verify the mechanism mechanism, models, data and application results, etc., shall not set up inducing user obsessed, excessive consumption, etc. Algorithm model. "As the technical logic of the platform's underlying technical logic, the algorithm becomes a direct supervision object. The subjective intention of the algorithm design deployment is expected to become the basis for accountability. The first public data utilization case tried by the Hangzhou Internet Court, the dispute originated from the "plaintiff bankruptcy liquidation" information without review. Although the information is automatically extracted and generated by the algorithm, the actual operating company of the platform should have certain attention obligations in the process of algorithm design, operation supervision, and results review. "Technical neutral" is not the same as "value neutral". The technology itself is neutral, but after being injected into the design concept, it is responsible. Therefore, on the basis of the platform design, deployment, and application of the platform design, deployment and application of the algorithm recommendation service provider, comprehensively consider the actual control of different infringement behaviors, predictability, subjective intentions, and afterhand relief measures, etc. The principle of blame for equivalent algorithms, promote platform autonomy to develop and improve within the framework of the rule of law. For example, for the recommendation of artificial intelligence media algorithms, the platform that requires the recommendation service provider of the algorithm can continuously improve the algorithm operation mode, reconcile the relationship between the trend algorithm and personalized algorithm with more closely related to the public discourse, and reduce the "information cocoon house" while providing services. effect.

Refine regulatory classification standards

The "Regulations" shows the legislative concept of classification and grading and accurate governance in regulatory ideas. On the one hand, according to public opinion attributes or social mobilization capabilities, content categories, user scale, data important degree of data processing of technical processing, intervention of user behavior, etc., on the other hand The provider makes specific specifications.

The "Provisions" proposes the concept of "the algorithm recommendation service provider with public opinion attributes or social mobilization capabilities". Such as information filing procedures and more stringent obligations such as security assessment. According to the algorithm recommendation service provider's power, user scale, etc., the corresponding class -level management mechanism is conducive to better achieve layered governance. However, the connotation and logical boundary of the "public opinion attributes or social mobilization capabilities" still need to be further clarified in judicial practice to avoid blurring areas. Including the nature conversion of different algorithm recommendation service providers in the application scenarios, and also the effective connection of the enhanced classification management system and the content of the algorithm content, and dynamic adjustment management. This should be the innovative space left by legislation for further expansion of judicial expansion. In the layered governance of the judicial protection objects, the "Regulations" made a more targeted specific specification for the service providers provided by minors, elderly, workers, consumers and other subjects. Emphasizing the induction of minors to avoid algorithm recommendation services, and pay attention to the safety and convenience of the elderly's recommendation services. How to develop the model recommendation service service for minors, elderly and other subjects? An app reviewed by the Hangzhou Internet Court to infringe the public welfare case of minors' network rights protection public welfare cases. In the trial, the court found that the APP may have a design that endangered the personal information rights and privacy of the public in the general unspecified minor's personal information and privacy in the collection, storage, use, and sharing personal information. Therefore, it not only ordered the platform to compensate for the loss of social public interests and apologies, but also urged them to block and disconnect a series of rectification tasks such as the recommendation service of privacy algorithms involved in minors. At the same time, leading the establishment of a minor network protection public welfare fund, etc., providing the judicial system foundation for strengthening the protection service protection of minors. Minor netizens are huge, which is of great significance to ensure the safety of minors and peaceful life. Through the above cases, we can see that it is necessary to further strictly collect and use the information of minor information, build a special review and human property guarantee mechanism for the recommendation service of minor algorithms, and effectively build the legislation of the governance of the algorithm of minors. Assure.

Adhere to the concept of digital power driving

In a new social pattern of networking, digitalization, and intelligence, algorithms will become an important factor affecting people's ideological decisions and shaping social concepts. This time the "Regulations" set the "algorithm see the door to see the door" the attention obligations and user equity protection of the algorithm recommendation service provider. From the aspects of algorithm recommendation technology, information content management, recommendation page display, specific user objects, etc. Essence However, in addition to strengthening the regulations at the level of establishment of the "Regulations", it is also necessary to highlight the value pursuit of "digital power" in the governance of judicial rules. The dynamic legal interest balance.

For example, the protection level of digital human rights and balance of legal benefits. From the perspective of judicial practice, some takeaway platforms have excessive pursuit of food delivery efficiency to meet their own economic benefits and user equals in the design of order distribution, compensation composition and payment, working hours, rewards and punishments. Brother "Labor Rights Protection and other digital human rights issues. Another example is the platform's independent management right and the algorithm recommendation service of the algorithm behind the user's independent option. The first -time consumer and e -commerce platform retrieval service disputes tried by the Hangzhou Internet Court determined that search services for e -commerce platforms can be lower than search engine standards to "precise matching". This not only fully considers consumers' dependence on platform recommendations and then strictly review the rationality of the algorithm, but also respects the platform's independent management rights.

The "Provisions" system gives users the right to run the system for choosing the right to choose, the right to know, the explanation and relief, and the relief system to run the full cycle of operation. It allows individuals to fight algorithm control through the "number of rights" from the legal mechanism. Diverse. In addition to the current provisions, you can consider trying to embed the social evaluation mechanism and expert technical ethics review mechanism, create opportunities to allow the public to effectively participate in the deployment of platform algorithm system, evaluate the consequences of the application algorithm, and promote the platform algorithm structure based on public value orientation. optimization.

(The author is the secretary of the party group and dean of the Hangzhou Internet Court)

Source: China Social Science Network-Journal of Social Sciences of China

Author: Hong Xuejun

- END -