Viewpoint | Analysis of Legal Issues in the Regulation of Platform Algorithms with Social Mobilization Capability-From the British Riots and the Influence of Social Platform Algorithms
Published:
2024-08-20
Recently, there have been protests in many parts of the UK around the government's immigration policy. Protests have spread to London, Manchester, Liverpool and other major cities, and in some areas have turned into violent riots, the largest violent riots in Britain in 13 years. In the meantime, Musk, the owner of the social media X platform, said: "The English Civil War is inevitable". The British government has blamed social media platforms for stoking tensions, especially Musk's X, and is considering early implementation of the Online Safety Act. Starmer said he was opposed to social networks "engaging in irresponsible behaviour", inciting violence and spreading disinformation. The British government will strengthen the supervision of online speech as the next focus to curb the source of violent unrest.
1. UK riots see social mobilization on social platforms
Recently, there have been protests in many parts of the UK around the government's immigration policy. Protests have spread to London, Manchester, Liverpool and other major cities, and in some areas have turned into violent riots, the largest violent riots in Britain in 13 years. In the meantime, Musk, the owner of the social media X platform, said: "The English Civil War is inevitable". The British government has blamed social media platforms for stoking tensions, especially Musk's X, and is considering early implementation of the Online Safety Act. Starmer said he was opposed to social networks "engaging in irresponsible behaviour", inciting violence and spreading disinformation. The British government will strengthen the supervision of online speech as the next focus to curb the source of violent unrest.
Of course, the root cause of the riots lies in the deep-seated reasons such as immigration contradictions, economic recession, social security, etc., but it cannot be ignored that social platforms with social mobilization capabilities do have considerable influence. Researchers at the University of Southern California (USC) found that the main culprit in the spread of misinformation is the incentive structure of the platform itself, rather than individual user behavior, with Facebook algorithms designed to prioritize sharing and rewarding inflammatory content. At the same time, we must see that in addition to artificial false information, generative artificial intelligence and deep synthesis technology are more threatening. The artificial intelligence model of deep learning and neural network algorithm, because it is an artificial simulation, does not have human ethical understanding of good and evil at all. It does not understand the input content, nor can it explain its generation mechanism and generation results. Once it is contaminated by training data and materials, the probability of generating illegal and bad content is extremely high, and the rapid spread caused by its viral transmission is an unavoidable problem in public governance in modern society.
What is 2. platform social mobilization capacity
In the early PC era, the ability of social mobilization was manifested in various forums, chat rooms, blogs, communication groups and other information services, providing channels for the expression of public opinion or Internet information services with the ability to mobilize the public to engage in specific activities. In the era of handheld intelligent terminals, it is manifested as information "cocoon rooms" caused by algorithms of various platforms themselves. These information "cocoon rooms" formed by interests and hobbies make a large number of public with consistent understanding gather with each other, constantly strengthen their cognition and understanding of something, and mobilize and coordinate actions with each other to produce certain social behaviors, this kind of social behavior may be positive (such as donation or rescue) or negative (such as rumors and riots), which is the social mobilization ability of the platform.
Modern society is constantly emphasizing and improving the ability of social governance. We are glad to accept the positive role of platform social mobilization ability, but we should also guard against the damage caused by improper information whirlpool. Therefore, the platform algorithm with the ability of social mobilization needs to be regulated by law.
How to carry out legal supervision on 3. platform algorithm with social mobilization ability
On May 21, 2024, the Council of the European Union approved the Artificial Intelligence Act (AI Act), which is the world's first special legislation for artificial intelligence. The Act divides artificial intelligence systems into four categories based on their potential impact on users and society: unacceptable risk category, high risk category, limited risk category, and minimum risk category. Each category applies different levels of regulatory requirements. Any AI system that attempts to manipulate human behavior, exploit human weaknesses or support government social scores would be considered an unacceptable risk and strictly prohibited. AI systems that pose a significant threat to health, safety, fundamental rights and the rule of law are classified as high risk. The Act requires such platforms to complete a pre-assessment-conformity assessment (Conformity Assessments)-before entering the market to confirm that they meet regulatory requirements related to transparency, data quality, record keeping, etc.
Similarly, my country has also promulgated and implemented corresponding laws and regulations in this field, such as the "Internet Information Service Algorithm recommend Management Regulations", "Internet Information Service Deep Synthesis Management Regulations", "Artificial Intelligence Interim Measures" and other regulations. How to achieve supervision?
(I) algorithm and model filing
After the algorithm and model are set, they can operate "autonomously", so as far as the platform is concerned, the algorithm and model itself is the direct object of supervision, and the primary means of regulation is filing. From the point of view of the record is not an administrative license, but the revocation of the record will be accompanied by the penalty of ordering the suspension of information update, enough to affect the operation of the platform.
An algorithm recommend service provider with the attribute of public opinion or the ability of social mobilization shall, within ten working days from the date of providing the service, fill in the service provider's name, service form, application field, algorithm type, algorithm self-evaluation report, content to be publicized and other information through the Internet information service algorithm filing system, and perform the filing procedures. The algorithm recommend service provider that completes the filing shall indicate its filing number in a prominent position on its website, application, etc. that provides services to the outside world and provide a link to the public information.
If an algorithm recommend service provider with the attribute of public opinion or the ability of social mobilization obtains the record by concealing the relevant information or providing false materials, the network information department of the state, province, autonomous region or municipality directly under the central government shall cancel the record, give a warning and circulate a notice of criticism; if the circumstances are serious, it shall be ordered to suspend the information update and impose a fine of not less than 10000 yuan but not more than 100000 yuan. If an algorithm recommend service provider with the attribute of public opinion or the ability of social mobilization terminates the service and fails to go through the formalities of cancellation and filing in accordance with the regulations, or is ordered to close the website, revoke the relevant business license or revoke the business license in a serious illegal situation, the network and information departments of provinces, autonomous regions and municipalities directly under the Central Government shall cancel the record.
(II) Algorithm Security Assessment
For example, if deep synthesis service providers and technical supporters provide models, templates and other tools with the following functions in the second paragraph of Article 15 of the regulations on the administration of deep synthesis of Internet information services, they shall carry out security assessment on their own or entrust professional institutions according to law: (1) generating or editing biometric information such as face and voice; (II) generating or editing non-biometric information such as special objects and scenes that may involve national security, national image, national interests and social public interests. (III) deep synthesis service providers to develop online new products, new applications, and new functions with public opinion attributes or social mobilization capabilities, safety assessments shall be carried out in accordance with relevant national regulations. Algorithm service providers, whether they provide recommend services, in-depth synthesis services or generative artificial intelligence services, should carry out security assessment in accordance with "relevant national regulations" if they "have public opinion attributes or social mobilization capabilities". The external manifestation of this security assessment is algorithm technology and its implementation methods, but the judgment standard is legal consequences, and its essence is platform algorithm compliance.
Algorithm security assessment should focus on the legality of information services and new technologies and new applications, the effectiveness of security measures stipulated in laws, administrative regulations, departmental rules and standards, and the effectiveness of prevention and control of security risks. The focus is on the following nine aspects:
1. Determine the person in charge of security management, information reviewers, or the establishment of a security management agency that is compatible with the services provided, and judge the adaptation of the agency's personnel, depending on the implementation of laws such as the Cyber Security Law, the Data Security Law, and the Personal Information Protection Law. The degree of regulation;
2. The user's real identity verification and registration information retention measures, whether through third-party platform authentication or direct registration, platform users are anonymous to each other, but for the platform, it is real name verification;
3. The user's account number, operation time, operation type, network source address and destination address, network source port, client hardware characteristics and other log information, as well as user release information records retention measures to ensure traceability;
4. The prevention and disposal of illegal and harmful information and relevant record-keeping measures in the service functions of platform user account name, nickname, introduction, remarks, identification, information release, forwarding, comment, etc., with the purpose of non-proliferation and traceability;
5. Algorithm recommend service providers shall regularly review, evaluate and verify the mechanism, model, data and application results of the algorithm, and shall not set up algorithm models that induce users to indulge in laws and regulations or violate ethics, such as excessive consumption, and implement the main responsibility of algorithm security;
6. Personal information protection and technical measures to prevent the spread of illegal and harmful information, social mobilization function out of control risk, once the risk is eliminated or reduced immediately;
7. Establish a complaint and reporting system, publish contact information, accept and handle relevant complaints and reports in a timely manner, and pay attention to the effective operation of the complaint reporting system;
8. Establish a working mechanism to provide technical and data support and assistance for the Internet information department and data security department to perform the supervision and management responsibilities of Internet information services in accordance with the law;
9. Establish a working mechanism to provide technical and data support and assistance for public security organs and national security organs to safeguard national security and investigate and deal with crimes in accordance with the law.
After safety assessment, in line with laws, administrative regulations, departmental rules and standards, should form a safety assessment report, safety assessment conclusions, and submitted to the relevant departments for the record.
The (III) algorithm model directly produces the content identification obligation.
In order to distinguish between "virtual reality" and "natural reality", it is necessary to distinguish between its generated content and objective facts. According to the "Regulations on the Administration of Deep Synthesis of Internet Information Services" jointly issued by the State Cyberspace Administration of China, the Ministry of Industry and Information Technology, and the Ministry of Public Security, "if a deep synthesis service provider provides deep synthesis services other than those specified in the preceding paragraph, it shall provide a significant identification function and prompt users of deep synthesis services to make significant identification." And the aforementioned identification of any organization and individual shall not use technical means to delete, tamper with, or conceal the depth of the composite logo. That is to say, the dissemination of such information must be accompanied by a distinctive identification, free from confusion in order to set the record straight.
The "Generative Artificial Intelligence Service Content Identification Method" (TC260-PG-20233A) issued by the National Information Security Standardization Technical Committee gives suggestions on the identification method, suggesting that the text content should at least contain information such as "artificial intelligence generation" or "AI generation", and the name of the service provider should be prompted by adding implicit watermark identification.
The aforementioned regulation is only a preliminary legal regulatory measure, data compliance is dynamic and constantly evolving, technology is constantly iterating, regulatory measures are also constantly changing, all changes remain the same, security is fundamental. The purpose of these regulations is to implement the relevant content of the Cyber Security Law, the Data Security Law, and the Personal Information Protection Law, especially for platforms with social mobilization capabilities, they should pay more attention to data security, social security, national security and economy. Security's legal obligations.
Key words:
Related News
Zhongcheng Qingtai Jinan Region
Address: Floor 55-57, Jinan China Resources Center, 11111 Jingshi Road, Lixia District, Jinan City, Shandong Province
