Promulgation Date: 2021-9-17 Title: Guiding Opinions on Strengthening the Overall Governance of Internet Information Service Algorithms. Document Number:国信办发文〔2021〕7号 [Source]http://www.cac.gov.cn/2021-09/29/c_1634507915623047.htm Promulgating Entities:State Internet Information Office, Central Propaganda Department, Ministry of Education, Ministry of Science and Technology, Ministry of Industry and Information Technology, Ministry of Public Security, Ministry of Culture and Tourism, State Administration for Market Regulation, State Administration of Radio and Television
In recent years, Internet information service algorithms (hereinafter referred to as "algorithms") have played an important role in areas such as accelerating the dissemination of information on the Internet, the prosperity of the digital economy, and promoting social development. At the same time, the unreasonable application of algorithms has also impacted the normal order of dissemination, market order, and social order, and brought challenges to the preservation of ideological security, social fairness and justice, and the lawful rights and interests of netizens. The following opinions on strengthening the security and governance of Internet information service algorithms are hereby issued in order to thoroughly implement the major decisions and arrangements of the Party Central Committee and the State Council, to manage the use of algorithms and develop the application of algorithms, and to fully improve the overall capacity for network governance.
I. General Requirements
(1) Guiding Ideology
Adhere to the guidance of Xi Jinping Thought on Socialism with Chinese Characteristics for a New Era, especially General Secretary Xi Jinping's important thinking on being a cyberpower; thoroughly implement the spirit of the 19th Party Congress and its 2nd, 3rd, 4th, and 5th, plenary sessions; persist in having positive energy as the general requirement, management as the final word, and good use as the real skill; orient towards the secure and reliable, high-quality, and innovative development of algorithms to establish and complete mechanisms for governing algorithm security, build and improve regulatory systems for algorithm security, advance algorithms self-innovation, and promote the healthy, orderly, and prosperous development of algorithms, to provide strong support for the development of a nation that is a cyberpower.
(2) Basic principles
Persist in the correct orientation, strengthening awareness of scientific ethics, security, and the bottom lines, giving full play to the role of algorithms in disseminating positive energy, and creating a clean and positive cyberspace; persist in governance in accordance with law, strengthening the establishment of laws and regulations, innovating models of technological oversight, combatting conduct that violates laws and regulations, and establishing and completing mechanisms for governing the security of algorithms in which multiple parties participate; persist in risk prevention and control advance the management of algorithim security by level and category and effectively identify high-risk algorithms to carry out precise governance; persist in safeguarding rights and interests, guide the use of algorithms to become fair and just, transparent and explainable, fully safeguarding the lawful rights and interests of netizens; persist in technical innovation, forcefully advancing our nation's research efforts on algorithm innovations, protecting intellectual property rights in algorithms, strengthening the deployment and spread of algorithms we researched, and increasing the core competitiveness of out nation's algorithms.
(3) Main objectives
Use about three years time to gradually establish an overall governance structure for algorithm security that completes governance mechanisms, improves oversight systems, and standardized the algorithm ecology.
--Complete governance mechanisms. Draft and improve policies and regulations on internet information service algorithm security with clear rights and responsibilities of entities in the governance of algorithm security and a government structure that operates at high efficiency, to form a governance mechanism that has law to rely on, coordinates multiple variables, and has the participation of multiple parties.
--Improved oversight systems. Innovatively construct and form an oversight system that integrates multiple functions such as algorithm security risk monitoring, algorithm security assessments, technological ethics reviews, algorithm filing and management, and handling of conduct involving algorithms that violates laws and regulations.
--Standardization of the algorithm ecology. Algorithms are to be correctly oriented with abundant positive energy, the use of algorithms is to be fair and equitable, the development of algorithms is to be safe and controllable and independently innovative, to effectively prevent the potential threats brought on by the abuse of algorithms.
II. Complete mechanisms for governing algorithm security
(1) Strengthen the standardization of the governance of algorithms. Complete policies and regulations on the governance of algorithm security, accelerate the drafting of provisions on managing algorithms, clarify the entities managing algorithms and the scope, requirements, and legal responsibility for the management; improve measures for the governance of algorithm security, draft standards, guidelines, and other accompanying documents.
(2) Optimize structures for the governance of algorithms. Further clarify the rights, obligations, and responsibilities of the government, enterprises, industry organizations, and netizens in the governance of algorithm security; scientifically and reasonably arranging the organizational structure for governance, standardizing operations and making interconnectivity, creating and forming a diverse and jointly governing overall layout for algorithm security with government regulation, enterprise fulfillment of responsibilities, industry self-discipline, and societal oversight.
(3) Strengthen overall planning and coordination for governance. In collaboration with departments such as for publicity, education, science and technology, industry and information, public security, culture and tourism, market regulation, and radio and television, the internet information departments are to establish a long-term and effective mechanism for coordinating and connecting departments, perform regulatory duties, and jointly carry out efforts to govern algorithm security.
(4) Strengthen enterprise entity responsibility. Enterprises should establish responsibility systems for algorithm security and for technology ethics reviews; complete organizational bodies for algorithm security management, strengthen governance of risk prevention and control and elimination of potential threats, and increase the ability to respond and level of response to emergency incidents in algorithm security. Enterprises should strengthen awareness of responsibility and bear entity responsibility for the results of applying algorithms.
(5) Strengthen industry organization self-discipline. The internet information service industry shall strengthen industry self-discipline, actively carrying out efforts to popularize algorithm science and technology, gradually putting together forces for the governance of algorithm security, attracting a professional talent pool, aggregating resource inputs from multiple areas, taking social responsibility for algorithm security governance, and providing strong supporting for the governance of algorithm security.
(6) Advocate netizen participation in oversight. Widely encourage netizens to actively participate in efforts on the governance of algorithm security, truly strengthening information exchanges and effective communication between government, enterprises, industry organizations, and netizens. Governments are to actively accept reports and complaints from netizens, and enterprises are to conscientiously accept societal oversight, and promptly give feedback on the outcomes.
III. Construction of a regulatory system for algorithm security
(7) Effectively monitor algorithm security risks. Carry out routine monitoring of algorithms' data usage, application scenarios, impact, and efficacy to perceive information such as trends in online transmission, changes in market rules, and netizen behavior brought on by the use of algorithms; get early warnings on potential threats such as irregularity, unfairness, or inequity that might be caused by the use of algorithms; and discover security problems in the use of algorithms.
(8) Actively carry out security assessments of algorithms. Organize and establish professional technical assessment teams to thoroughly analyze algorithm mechanisms, assess flaws and vulnerabilities in algorithm design, deployment, usage, and other phases of their application; assess security risks caused by the use of algorithms such as in terms of ideology, social fairness, morality and ethics, and submit focused response measures.
(9) Orderly advance efforts on the recording of algorithms. Establish a recording system for algorithms, cleaning up the basic circumstances for recording algorithms, completing systems for sorting algorithms by level and category, clarify the scope of recording for algorithms, and orderly implement efforts on recording. Actively complete guidance and assistance in recording, proactively publishing the circumstances of recording and accepting societal oversight.
(10) Continue to advance innovations in regulator models. Continue to assess new situations in the technical development of algorithms, advancing the coordinated development of regulatory models and algorithm technology, continuously improving, upgrading, and innovating regulatory means and methods and governance measures to prevent outmoded regulatory models causing algorithm security risks.
(11) Strictly combat conduct that violates laws and regulations. Strive to resolve issues in algorithm security on which netizens give strong feedback, strictly combat conduct involving algorithms that violates laws and regulations that is discovered in the course of monitoring, assessment, and recording work, or that is reported by netizens and verified; persist in preserving the security of internet information service algorithms.
IV. Promote the standardized development of the algorithm ecology.
(12) Establish the correct orientation for algorithms. Carry forward the Core Socialist Values, persisting in the correct political orientation, public sentiment orientation, and values orientation in the application of algorithms. Increase the precision and efficacy of disseminating positive energy, regulate conduct and order in information publication, promote enterprises' strengthening the dissemination of positive energy using algorithms, guiding the use of algorithms for good.
(13) Promote openness and transparency in algorithms. Regulate enterprises' conduct in the use of algorithms, protect netizen's reasonable rights and interests, uphold the principles of fairness and justice, and promote openness and transparency in algorithms. Urge enterprises to disclose information on algorithms' basic principles, optimization goals, decision-making standards in a manner that is timely, reasonable, and effective; do a good job of explaining algorithm results, clear channels for complaints, eliminate societal misgivings, and promote the healthy development of algorithms.
(14) Encourage innovation and development of algorithms. Increase capacity for innovation in algorithms, actively carrying out research and development efforts on algorithms, supporting the deep integration of algorithms and society, and all areas of economics. Increase capacity for self-control of our algorithms, strengthening protections of intellectual property rights, increasing the spread and usage of algorithm products researched by ourselves, enhance core competitiveness in algorithms.
(15) Prevent risks of algorithm abuse. Preserve the order of cyberspace transmission, market order, and societal order; prevent the use of algorithms for conduct such as interfering with societal public opinion, suppressing competitors, or infringing on netizens' rights and interests; and prevent potential threats in areas such as ideology, economic development, and social management that might be brought on by the abuse of algorithms.