This page is still being polished. If you have thoughts, please share them via the feedback form.
Data on this page is preliminary and may change. Please do not share or cite these figures publicly.
Internal policies, content safety guidelines, and ethical design principles governing system creation.
Also in Engineering & Development
Reasoning
Safety guidelines establish ethical design principles governing model algorithm developer practices.
Developers should uphold a people-centered approach, adhere to the principle of AI for good, and follow science and technology ethics
Developers should uphold a people-centered approach, adhere to the principle of AI for good, and follow science and technology ethics in key stages such as requirement analysis, project initiation, model design and development, and training data selection and use, by taking measures such as internal discussions, organizing expert evaluations, conducting technological ethical reviews, listening to public opinions, communicating and exchanging ideas with potential target audience, and strengthening employee safety education and training.
2.4.2 Design StandardsDetailed test reports
Developers should generate detailed test reports to analyze safety and security issues, and propose improvement plans.
2.2.4 Assurance DocumentationData security
Developers should strengthening data security and personal information protection, respect intellectual property and copyright, and ensure that data sources are clear and acquisition methods are compliant. Developers should establish a comprehensive data security management procedure, ensuring data security and quality as well as compliant use, to prevent risks such as data leakage, loss, and diffusion, and properly handle user data when terminating AI products
2.1.3 Policies & ProceduresSecure training environment
Developers should guarantee the security of training environment for AI model algorithms, including cybersecurity configurations and data encryption measures
1.2.4 Security InfrastructureAssess biases
Developers should assess potential biases in AI models and algorithms, improve sampling and testing for training data content and quality, and come up with effective and reliable alignment algorithms to ensure risks like value and ethical risks are controllable
2.2 Risk & AssuranceEvaluate readiness
Developers should evaluate the readiness of AI products and services based on the legal and risk management requirements of the target markets.
2.2.1 Risk AssessmentManage versions
Developers should effectively manage different versions of AI products and related datasets. Commercial versions should be capable of reverting to previous versions if necessar
2.4.3 Development WorkflowsSafety and security evaluation tests
Developers should regularly conduct safety and security evaluation tests. Before testing, they should define test objectives, scope, safety and security dimensions, and construct diverse test datasets covering all kinds of application scenarios.
2.2.2 Testing & EvaluationTest rules and methods
Developers should formulate clear test rules and methods, including manual testing, automated testing, and hybrid testing, and utilize technologies such as sandbox simulations to fully test and verify models.
2.2.2 Testing & EvaluationEvaluate tolerance
Developers should evaluate tolerance of AI products and services for external interferences and notify service providers and users in forms of application scope, precautions, and usage prohibitions.
2.2.2 Testing & EvaluationTechnological measures to address risks
Responding to the above risks, AI developers, service providers, and system users should prevent risks by taking technological measures in the fields of training data, computing infrastructures, models and algorithms, product services, and application scenarios.
1 AI SystemTechnological measures to address risks > Addressing AI’s inherent safety risks
99 OtherTechnological measures to address risks > Addressing safety risks in AI applications
99 OtherComprehensive governance measures
2.1 Oversight & AccountabilityComprehensive governance measures > Implement a tiered and category-based management for AI application
We should classify and grade AI systems based on their features, functions, and application scenarios, and set up a testing and assessment system based on AI risk levels. We should bolster enduse management of AI, and impose requirements on the adoption of AI technologies by specific users and in specific scenarios, thereby preventing AI system abuse. We should register AI systems whose computing and reasoning capacities have reached a certain threshold or those are applied in specific industries and sectors, and demand that such systems possess the safety protection capacity throughout the life cycle including design, R&D, testing, deployment, utilization, and maintenance.
3.1.4 Compliance RequirementsComprehensive governance measures > Develop a traceability management system for AI services
We should use digital certificates to label the AI systems serving the public. We should formulate and introduce standards and regulations on AI output labeling, and clarify requirements for explicit and implicit labels throughout key stages including creation sources, transmission paths, and distribution channels, with a view to enable users to identify and judge information sources and credibility.
3.1.4 Compliance RequirementsAI Safety Governance Framework
National Technical Committee 260 on Cybersecurity of SAC (2024)
Artificial Intelligence (AI), a new area of human development, presents significant opportunities to the world while posing various risks and challenges. Upholding a people-centered approach and adhering to the principle of developing AI for good, this framework has been formulated to implement the Global AI Governance Initiative and promote consensus and coordinated efforts on AI safety governance among governments, international organizations, companies, research institutes, civil organizations, and individuals, aiming to effectively prevent and defuse AI safety risks.
Build and Use Model
Training, fine-tuning, and integrating the AI model
Governance Actor
Regulator, standards body, or oversight entity shaping AI policy
Govern
Policies, processes, and accountability structures for AI risk management
