Responsible Artificial Intelligence Policy Statement
Sinoma International Engineering Co., Ltd. (hereinafter referred to as the Company) strictly abides by the Personal Information Protection Law of the People's Republic of China, the Data Security Law of the People's Republic of China, the Cybersecurity Law of the People's Republic of China, the Interim Measures for the Administration of Generative Artificial Intelligence Services, and other laws, regulations, and normative documents concerning data protection, algorithm governance, and cybersecurity that have a material impact on the Company. This Policy Statement is also informed by international standards and frameworks, including the OECD Principles on Artificial Intelligence, the UNESCO Recommendation on the Ethics of Artificial Intelligence, and the EU Artificial Intelligence Act (EU AI Act). This Policy Statement is subject to overall oversight by the Board of Directors, jointly formulated by the Board and relevant senior management, and is subject to regular review and improvement.
This Policy Statement applies to the development, deployment, and application of artificial intelligence technologies across the Company's headquarters and its subsidiaries during production and operations, R&D and innovation, procurement and sales, logistics and transportation, customer services, and other activities. It covers scenarios including intelligent production control, intelligent quality inspection, smart mining, smart logistics, as well as relevant collaboration with third parties such as suppliers, contractors, and technology partners.
I. Data Privacy and Cybersecurity
1. Strictly comply with applicable data protection laws and regulations, ensuring that data used for AI training and application is lawfully sourced, has clear ownership, and is fully authorized.
2. For core sensitive production data, implement the highest level of protection in accordance with data classification and grading principles, and strictly control the scope of use of such data in AI training and inference.
3. Adopt technical measures including data encryption, desensitization, anonymization, and access control to prevent data leakage, misuse, or unauthorized access, thereby effectively safeguarding the data rights and interests of customers, partners, and employees.
4. When deploying AI in industrial control systems (DCS/SCADA), implement stringent cybersecurity isolation and defense-in-depth measures.
5. Implement full-lifecycle security management for AI systems throughout the entire lifecycle — from project initiation, development, and deployment to operations and maintenance. Gradually establish and enhance security testing and risk assessment mechanisms, ensure training data quality, and periodically verify system robustness and security to mitigate potential risks.
6. Ensure that the AI operating environment is secure, stable, and controllable, and prevent production incidents or data breaches resulting from cyberattacks or system failures.
7. Strictly restrict access to sensitive AI capabilities. Use cases involving facial recognition and biometric monitoring are subject to special approval, with risk assessment and compliance review completed prior to deployment to ensure compliance with applicable laws, regulations and international norms.
II. Algorithmic Fairness and Bias Mitigation
1. In AI technology application scenarios, proactively identify and mitigate bias in training data, and build diverse and representative datasets.
2. Ensure that AI systems do not produce discriminatory outcomes based on factors such as age, gender, or physical ability in the course of service delivery, decision-making, or recommendation.
3. Establish and regularly implement a fairness assessment mechanism for deployed AI models. Through algorithm audits, bias‑detection tools and multi‑dimensional tests, the Company continuously monitors the impartiality of model outputs and promptly rectifies any identified biases.
4. Establish mechanisms to detect and rectify drift or degradation of AI models over time, and periodically review and retrain models for performance, accuracy and stability to ensure the consistent reliability of model outputs.
III. Human Involvement and Intervention in Critical Decisions
1. In scenarios involving physical safety and major production operations, adhere to the "human-in-the-loop" principle.
2. Maintain human involvement at critical stages and establish human intervention mechanisms. AI shall serve as an assistive tool, with necessary human review and ultimate decision-making authority preserved.
3. Provide clear labelling and explicit disclosure for AI‑generated content, AI‑driven decision‑making outcomes and their impacts, enabling users and relevant stakeholders to accurately distinguish AI outputs from human‑led judgements.
IV. Transparency and Explainability
1. Clearly communicate the use cases, capability boundaries, and potential risks of AI systems, and ensure that users and employees are aware when they are interacting with an AI system.
2. For the various determinations and alerts generated by AI systems, provide understandable rationale in tandem with the output; outputs shall not be limited to scores or conclusions alone.
3. Strictly define the scope of AI application across business scenarios, clearly specify the tasks AI may perform and those for which it is unsuitable or prohibited, and prevent technology misuse or out-of-scope application.
V. AI Accountability and Capability Boundaries
1. Establish a responsible AI governance framework spanning R&D, deployment, operations, and decision-making, with clearly defined accountability for each stage.
2. Gradually establish relevant processes for investigating and addressing errors or adverse impacts that may arise from AI systems, ensuring that risks and legal responsibilities associated with AI-driven decisions are effectively identified and managed.
VI. Green and Sustainable AI Development
1. Monitor the environmental impact of both in-house and third-party AI infrastructure, and encourage the use of low-carbon, energy-efficient, and water-efficient data centers, green computing models, and renewable energy.
2. Actively promote the application of AI technologies in energy conservation, emission reduction, green and safe production, prioritize the adoption of computing facilities with high energy‑efficiency ratings, collaborate with suppliers to advance green computing solutions, and regularly evaluate and quantify environmental benefits for incorporation into the project assessment system.
VII. Prohibition of High-Risk and Improper AI Applications
1. The Company shall never develop or deploy AI systems designed for purposes such as manipulating human behavior, exploiting the vulnerabilities of specific groups, conducting social scoring, or performing unauthorized biometric identification. The Company will continue to monitor international AI governance trends and comply with requirements concerning high-risk AI applications prohibited by international norms.
2. Strictly prohibit the deployment of manipulative, exploitative or discriminatory AI systems to ensure all AI applications align with social ethics and corporate values.
VIII. Responsible AI Governance and Oversight Mechanism
1. The Company shall establish an AI governance and oversight mechanism comprising management, business leaders, and external experts, operating under the supervision and direction of the Board of Directors.
2. Continuously enhance the AI governance mechanism, conduct AI governance compliance reviews, strengthen risk management, and proactively accept oversight from regulatory authorities and the public.
3. Strengthen employee awareness and communication and enhance responsible AI awareness among internal and external stakeholders. For employees, we conduct AI ethics and safety training to foster understanding of the boundaries, risks, and rights associated with AI applications. For other stakeholders, we maintain open communication through multiple engagement channels to ensure that our suppliers, technology partners, and others are also aware of and fulfil their AI responsibilities. For the public, we are committed to promoting safe, compliant, and sustainable AI applications in the industry in a responsible manner.