Mrs. Li Hualan
| BIOGRAPHY
Mrs. Nancy Li has over 15 years of experience in cybersecurity and privacy, with extensive expertise in AI security, R&D security engineering, and security governance. Her areas of competence and expertise cover security and privacy across multiple domains, including AI security, 5G security, R&D security engineering, security governance, and privacy protection.
| TOPIC
Strengthening Infrastructure Security and Safeguarding All Intelligence
| ABSTRACT
Huawei’s AI Security deck presents the company’s perspective and technical approach to securing AI as the field shifts from Artificial Narrow Intelligence toward AGI and ASI. The document is organized in three parts. First, it maps ten categories of AI security and privacy threats across development and operations phases, including prompt injection, data poisoning, adversarial examples, model extraction, deepfakes, and task hijacking. Second, it surveys industry AI governance: international initiatives (UNESCO, OECD, Bletchley), divergent national approaches (EU AI Act, China’s generative AI rules, US deregulation), and standards bodies (ISO/IEC JTC 1/SC42, ETSI SAI). Third, it details Huawei’s response: six Responsible AI governance principles operationalized through five defense-in-depth guardrails along the foundation model lifecycle, anchored on four trustworthy fundamental technologies (Crypto Engine, Confidential Computing, Trusted Computing, Trusted Interconnection) and a situation-awareness O&M system, positioned as an independent, controllable alternative security standard.