AI Usage Statement

Last Updated: June 5, 2026 (Optimized Version 1.1)

This statement transparently explains the use of artificial intelligence technology on this platform, including data collection, processing, storage, sharing, and user rights, in accordance with the requirements of the Interim Measures for the Management of Generative Artificial Intelligence Services (effective August 15, 2023), the Provisions on the Administration of Algorithmic Recommendations in Internet Information Services (effective March 1, 2022), the Personal Information Protection Law of the People's Republic of China (PIPL), and the Data Security Law of the People's Republic of China (DSL).

I. AI Technology Application Scenarios

This platform uses artificial intelligence technology in the following scenarios. Each scenario follows the principle of minimal necessity and fulfills the obligation of notification in accordance with the law.

Application Scenario	AI Function	Data Input	Human Oversight
Intelligent Customer Service	Natural language understanding and answer generation	User dialogue content (including potential personal information)	Manual review of abnormal responses; continuous optimization of dialogue quality
Knowledge Base Retrieval	Semantic search and relevance ranking	Enterprise-uploaded documents and vectorized embeddings	Regular evaluation of retrieval results; users can manage the knowledge base themselves
Content Recommendation	Personalized recommendations based on user behavior	Browsing behavior data (de-identified), user profile tags	Review mechanism for recommendation results; users can disable at any time
Content-Assisted Generation	Generation of copy, summaries, and other content	User-provided materials and instructions	Generated content must be reviewed manually before publication; generation records are traceable
Data Analysis	Trend prediction and anomaly detection	Built-in business statistical data (excluding personal identity information)	Manual verification of analysis results; used for internal operations only

Details of data processing in each scenario are as follows:

• Intelligent Customer Service: AI models are invoked via API to process user dialogue content (including text and potential personal identifiers). Dialogue records are retained for 180 days for service response and quality optimization. If such content is to be used for model training in the future, we will obtain your separate consent (PIPL Articles 17 and 23). Dialogue content may be transmitted in real time to third-party AI service providers (e.g., large language model providers) for natural language processing. Encryption measures are adopted during transmission, and the data is deleted immediately after processing.
• Knowledge Base Retrieval: Enterprise documents you upload (e.g., PDF, Word) are stored in the knowledge base system, and vectorized embeddings are generated for semantic search. Documents and vector data are retained for the same period as the source documents (you may delete them at any time), and vectorized embeddings are deleted in sync with the source documents. To provide retrieval capabilities, document content may be shared with third-party AI service providers for vectorization processing (the process complies with the Data Security Law).
• Content Recommendation: Personalized recommendations are based on your browsing behavior (e.g., clicks, dwell time) and user profile tags (e.g., interests and preferences). Behavioral data and tags are de-identified and retained for 2 years. You can disable personalized recommendations in your account settings; after disabling, content with a default, non-personalized ordering will be displayed.
• Content-Assisted Generation: The materials and instructions you input are used to generate content via AI. Generation records (including original inputs and outputs) are retained permanently for version tracking and compliance auditing. These records may be shared with third-party AI service providers for real-time generation; the third parties do not store data after generation is complete. You have the right to request deletion of your generation records.
• Data Analysis: Only de-identified business statistical data (e.g., usage frequency, error rate) is used, and no personally identifiable information is involved.

II. Algorithm Filing and Transparency

Pursuant to Article 24 of the Provisions on the Administration of Algorithmic Recommendations in Internet Information Services:

• The algorithmic recommendation services we use (personalized recommendations based on user behavior and profiles) have been filed with the cyberspace administration department in accordance with the law.
• Filing number: [To be supplemented]
• Basic algorithm principles: Based on natural language processing and deep learning models, combined with users' short-term and long-term behavioral characteristics and content tags, semantic matching and relevance ranking are performed.
• Algorithm operation mechanism: User profiles are generated by analyzing users' browsing history, search keywords, and interaction preferences, and the recommendation ranking is dynamically adjusted.
• You can disable the personalized recommendation function at any time (path: Account Settings → Privacy → Personalized Recommendations); after disabling, default content ordering (e.g., by time or popularity) will be displayed.
• To provide personalized recommendations, the platform collects user behavior data (via cookies/tracking technologies) and user profile tags, and your consent has been obtained prior to collection (PIPL Article 13).

III. AI-Generated Content Labeling

Pursuant to Article 12 of the Interim Measures for the Management of Generative Artificial Intelligence Services:

• Content generated or assisted by AI will be labeled as "AI-generated" or "AI-assisted".
• AI-generated images will include invisible digital watermarks for provenance verification.
• Users can distinguish between human-created and AI-generated content through these labels.
• AI content generation records (including labeling information) are retained permanently to facilitate regulatory review.

IV. Data Usage, Model Training, and Third-Party Sharing

This platform is involved in the following data processing activities during the provision of AI services, as detailed below:

Data Type	Collection Method	Purpose	Retention Period	Shared with Third Parties	Sharing Object and Purpose	Legal Basis
AI dialogue content	User input via intelligent customer service interface; API transmission	Service response, quality optimization	180 days (automatically deleted)	Yes	Third-party LLM service provider (for real-time Q&A processing)	PIPL Article 13 (consent)
Content sent to LLM for processing (user instructions/materials)	Real-time API transmission	AI content generation and intelligent analysis	Deleted immediately upon processing (third party does not retain)	Yes	Third-party LLM service provider (for immediate processing)	PIPL Article 13 (consent); Article 30 of the Data Security Law
AI content generation records	Automatic system logging	Version tracking, quality auditing, compliance archiving	Permanent (users may request deletion)	Yes	Third-party AI service provider (for generation process; no copies retained)	Article 7 of the Interim Measures for Generative AI Management
Document uploads and knowledge base content	User upload	Enterprise knowledge management and RAG retrieval	Permanent (users may delete on their own)	Yes	Third-party vectorization/retrieval service provider	Article 21 of the Data Security Law; Interim Measures for Generative AI Management
Document vectorized embeddings	Automatically generated by AI	Semantic search and knowledge retrieval	Synchronized with source documents (deleted when source documents are deleted)	Yes	Third-party AI service provider (vectorization processing)	PIPL Article 13 (contract performance)
AI orchestration decision logs	Automatic system logging	Audit and explainability of AI autonomous decisions	2 years (automatically archived)	No	Not shared	AI Safety Governance Framework
Algorithm recommendation and personalization rules	System configuration and operation logs	Personalized content recommendations	Duration of rule validity	No	Not shared	Article 16 of the Algorithmic Recommendation Provisions
User profiles and behavior tags	Generated via behavior analysis models	Personalized recommendations and targeted marketing	2 years (de-identified)	No	Not shared	PIPL Article 13 (consent; can be disabled)
User behavior data samples (de-identified)	Collected via tracking technologies	Brand profile optimization and market insights	1 year (de-identified aggregation)	No	Not shared	PIPL Article 13 (legitimate interests)
AI model usage statistics	Automatic API statistics	Service monitoring and cost accounting	1 year	No	Not shared	PIPL Article 13 (legitimate interests)
User authentication and login credentials	Form registration/login	Identity verification and access control (AI service entry)	Duration of account existence	Yes	Third-party identity verification service provider (verification only)	Article 24 of the Cybersecurity Law (real-name system)

Note: For all personal information or important data shared with third parties, we have completed assessments and signed data processing agreements. For scenarios requiring separate consent (e.g., using dialogue content for model training), we will display a clear prompt during the first interaction and obtain your authorization.

V. AI Usage Limitations

• AI-generated content is for reference only and does not constitute professional advice in legal, medical, financial, or other fields.
• AI systems may produce inaccurate or biased responses; users should exercise rational judgment.
• Users should not rely on AI responses to make significant decisions.
• AI systems are equipped with content safety filtering mechanisms that refuse to generate illegal or non-compliant content (e.g., pornography, violence, discrimination), and undergo regular safety assessments.

VI. User Rights

• You have the right to choose not to use AI-related features, including disabling intelligent customer service and prohibiting personalized recommendations.
• You have the right to request deletion of your dialogue history with the AI system (path: Privacy Settings → Dialogue Record Management).
• You have the right to request deletion of your uploaded knowledge base documents and their vectorized data.
• You have the right to disable algorithmic recommendation features (path: Account Settings → Privacy → Personalized Recommendations).
• You have the right to file complaints and suggestions regarding AI services.
• We will respond within 15 business days after receiving your request (PIPL Articles 44 to 47).