"Basic Security Requirements for Generative Artificial Intelligence Services" is open for comment

IT Home News on October 12, according to the official website of the National Information Security Standardization Technical Committee, the technical document "Basic Requirements for Generative Artificial Intelligence Service Security" (hereinafter referred to as the "Requirements") organized by the committee was released on October 11 A draft for solicitation of opinions was formed on October 25, and the technical document is now open to the public for solicitation of opinions. If you have any comments or suggestions, please provide feedback before 24:00 on October 25.

The "Requirements" provide the basic security requirements for generative artificial intelligence services, including corpus security, model security, security measures, security assessment, etc. Applicable to the provision of generative artificial intelligence services to the public in my country The provider improves the security level of the service, or the provider conducts a security assessment on its own or entrusts a third party to conduct a security assessment. It can also provide relevant authorities with the ability to evaluate the security level of generative artificial intelligence services in order to rewrite it without changing the original intention. Content, the language needs to be rewritten to Chinese.

Image source Pexels

IT House’s organizing requirements are as follows:

• Establish a blacklist of corpus sources, and data from blacklist sources must not be used for training.
• Security assessments should be conducted on each source corpus. If the content of a single source corpus contains more than 5% of illegal and harmful information, it should be added to the blacklist.
• When using corpus containing personal information, the authorization and consent of the corresponding personal information subject should be obtained, or other conditions for the legal use of the personal information should be met.
• When using corpus containing biometric information such as faces, the written authorization and consent of the corresponding personal information subject should be obtained, or other conditions for the legal use of the biometric information should be met.
• The annotators should be assessed by themselves, those who pass the test should be given annotation qualifications, and there should be a mechanism for regular retraining and assessment, as well as a mechanism for suspending or canceling the annotation qualifications when necessary.
• During the training process, the security of the generated content should be one of the main considerations for evaluating the quality of the generated results.
• Those who provide services through an interactive interface should disclose the following information to the public in a prominent position such as the homepage of the website:

The content that needs to be rewritten is: -Applicable people, occasions, purposes and other information

-Third-party base model usage.

• The necessity, applicability and safety of applying generative artificial intelligence in various fields within the service scope should be fully demonstrated.

According to previous reports from IT House, seven departments including the Cyberspace Administration of China, the National Development and Reform Commission, the Ministry of Education, the Ministry of Science and Technology, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration of Radio and Television issued the "Interim Provisions on Generative Artificial Intelligence Service Management" on July 10. Measures", which will be implemented on August 15

The "Measures" stipulate that you must not take advantage of algorithms, data, platforms, etc. to implement monopoly and unfair competition; you must not endanger the physical and mental health of others, and you must not infringe other people's portrait rights, reputation rights, honor rights, privacy rights and personal information rights; Effective measures should be taken to enhance the transparency of generative artificial intelligence services and improve the accuracy and reliability of generated content.

In order to rewrite the content without changing the original meaning, the language needs to be rewritten to Chinese

The above is the detailed content of "Basic Security Requirements for Generative Artificial Intelligence Services" is open for comment. For more information, please follow other related articles on the PHP Chinese website!