With the vigorous development of the digital economy and the acceleration of the new generation of technological revolution, Artificial Intelligence (AI) is reshaping business logic and social forms in an unprecedented depth. While actively embracing AI technology and promoting business intelligence transformation, Yihai Kerry promises to strictly abide by relevant laws and regulations, clarify the company's ethical boundaries and action guidelines in the field of AI, and ensure that technological progress always goes in the same direction as safeguarding personal rights, enhancing public trust, and promoting environmental friendliness.

Careful Oversight and Management of Sensitive Artificial Intelligence Capabilities

The Company carefully uses sensitive artificial intelligence technologies such as image recognition and facial recognition in scenarios such as smart factories and warehousing logistics, which are mainly applied to equipment inspection, product quality testing, and production safety management. The Company clearly defines the usage boundaries, employee permissions, approval processes, and supervision mechanisms of relevant technologies, strictly limits them to the scope required for production safety and compliance, and does not use them for improper purposes such as employee behavior monitoring, in order to protect employees' individual rights and maintain internal and public trust.

Distinct labeling of AI-generated content and outcomes of AI-driven decisions

For marketing materials, intelligent customer service Q&A, and other content related to AI generation or auxiliary generation, relevant business departments of the Company will clearly label and explain, to help internal and external stakeholders clearly confirm the source of information provided (human-created or AI-generated). In important decision-making scenarios involving personnel management, AI is only used as an auxiliary analysis tool, and the final conclusions need to be manually reviewed and confirmed to ensure that the decision-making process is transparent and traceable.

Continuous monitoring and correction of AI model drift

For AI applications such as production quality inspection and sales forecasting, the Company has established a regular back testing and online monitoring mechanism for model performance, continuously tracking the deviation between model output and actual results. When there are signs of accuracy decline or model drift, the Company will promptly retrain the model with the latest data or adjust its parameters to ensure that the model maintains stable and reliable performance in the long term.

Regular assessments of deployed AI model for fairness and bias 

In AI-assisted applications related to personnel, the Company focuses on regular review of model outputs, paying attention to whether there are unreasonable differences caused by data bias, and calibrating model rules based on professional judgment from business and human resources departments, striving for AI-assisted decision-making to treat different age, gender, ethnicity, nationality, and other groups equitably and eliminate discriminatory bias.

Pay attention to the ecological footprint of AI data centers and models

The Company continues to promote the greening of IT infrastructure, reducing energy consumption of computing resources through server virtualization, cloud resource elastic scheduling, and data center energy efficiency optimization. In the AI model development process, emphasis is also placed on optimizing algorithms for lower energy consumption, and helping the company achieve overall energy conservation, emission reduction, and "dual carbon" goals.

Appeal channels for users or business-related parties to contest an AI decision or outcome

For scenarios involving AI-assisted results such as intelligent customer service responses and order risk alerts, the Company provides multi-channel manual appeal and review channels to ensure that customers, suppliers, and other relevant parties can receive timely human intervention and response when they have doubts about AI-driven decisions, and that the appeal process is accessible, responsive and effective.

Quantification of the impact of AI on sustainability outcomes

The Company is currently attempting to quantify and track the sustainable benefits of AI application scenarios such as intelligent production scheduling, energy consumption optimization, and logistics route planning, covering specific indicators such as energy consumption reduction and loss rate reduction, and gradually incorporating them into the Company's sustainable development management system. This data-driven approach aims to enhance the measurability and transparency of AI measures' contributions to environmental and social value.

Training of employees on the ethical use and security of AI

The Company continues to provide digital literacy and AI application training programs for employees, covering topics such as data privacy, information security prevention, and compliant use of AI, to help employees correctly understand the ability boundaries and potential risks of AI technology. The intelligent portal assistant "Xiaoyu AI" within the Company is a practical case of continuous iterative optimization under this concept, committed to serving employees' daily work in a safe, transparent, and reliable manner.