Select a Country or Region
- Australia - English
- Brazil - Português
- China - 简体中文
- Europe - English
- France - Français
- Germany - Deutsch
- Ireland - English
- Italy - Italiano
- Japan - 日本語
- Kazakhstan - Қазақ тілі
- Kazakhstan - Русский
- Kenya - English
- Korea - 한국어
- Malaysia - English
- Mexico - Español
- Mongolia - Монгол
- Netherlands - Nederlands
- New Zealand - English
- Poland - Polski
- Romania - Română
- Singapore - English
- South Africa - English
- Spain - Español
- Switzerland - Deutsch
- Switzerland - Français
- Switzerland - Italiano
- Switzerland - English
- Tanzania - English
- Thailand - ภาษาไทย
- Türkiye - Türkçe
- Ukraine - Українська
- United Kingdom - English
- Uzbekistan - Русский
- Uzbekistan - O’zbek
- Vietnam - Tiếng Việt
- Global - English
This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy
For decades, ICT talent development has rested on a basic assumption: Technology was relatively stable, and a single round of systematic learning could sustain a career. That assumption no longer holds. AI-native is becoming the new industry paradigm as large models evolve from basic tools into the very foundation of productivity. The shelf life of skills is now measured in months, yet traditional training cycles are still counted in years. This paradigm shift, one that is fundamentally about people, has already begun.
A paradigm shift: AI-native is rewriting the ICT job landscape
The shift is already visible in the job market. Traditional, standardized roles, such as basic network configuration, routine integration, and repetitive testing and maintenance, are gradually shrinking. At the same time, new roles are emerging rapidly: AI agent orchestration, retrieval-augmented generation (RAG) engineering, large language model (LLM) operations, and AI scenario deployment. The World Economic Forum's The Future of Jobs Report 2025 projects that by 2030, 39% of workers' existing skill sets will be disrupted, and 59% of workers will need reskilling. China's Ministry of Human Resources and Social Security has added roles like "Generative AI System Application Specialist" to its new occupation directory. The industry's talent needs are shifting from technical execution to value creation. What enterprises need is no longer engineers who can follow a plan, but multi-skilled talent who can use AI to generate business value.
Roles are evolving, and skills are depreciating faster than ever. The industry consensus is that the half-life of technical skills has shrunk to under five years, and for LLM-related skills, the renewal cycle is measured in months. Yet traditional training models are struggling with three major shortcomings. First, curricula cannot keep pace: Textbooks take years to develop, and by the time students graduate, the knowledge they have acquired is often already outdated. Second, hands-on practice is disconnected from the real world: Learners study AI through slides and learn about agent orchestration without ever actually orchestrating one. Third, assessment is too narrow: A written exam cannot measure practical abilities like human-AI collaborative development or deployment in complex scenarios. All of this is resulting in a growing talent gap between training supply and industry demand.
Breaking the "impossible trilemma": From static programs to dynamic closed-loop learning
Education has long faced an impossible trilemma: the difficulty of achieving scale, quality, and personalization all at once. One-on-one tutoring offers quality and personalization, but cannot scale. Traditional MOOCs achieve scale, but at the cost of personalization. Training without data is merely rote repetition of experience, and training without a closed loop is only a one-off effort.
The Shixizhi 2.0 AI Learning Platform is Huawei's digital and intelligent learning platform, and it is driven by both AI and data. The platform features a learning agent and an enterprise AI knowledge base that enable personalized, interactive, and granular learning experiences and professional exam services for employees, partners, ecosystem players, customers, and the public. It also integrates high-quality learning content into workflows to support the intelligent upgrades of business and the growth of both organizations and talent. The main purpose of Shixizhi 2.0 is to break this trilemma using the intelligent, data-driven, and dynamic capabilities of AI-powered learning platforms.
The platform is built on four core capabilities. AI-driven learning analytics track each learner's knowledge status and behavior, replacing experience-based course planning. A job competency map breaks down industry needs into measurable skill units, making proficiency quantifiable. Dynamic course updates keep content aligned with emerging technologies, preventing textbooks from becoming outdated before they are even used. Context-specific intelligent training brings real engineering environments into the classroom, allowing learners to learn by doing.
These four interconnected capabilities form a closed-loop system that links industry needs, skill breakdown, targeted training, and competency assessment (see Figure 1). Assessment results feed back into learning analytics, and industry changes update the competency map in real time. Talent development thus evolves from a static one-off effort into a dynamic continuously operating, self-improving system.
Figure 1: A closed-loop system that connects industry needs, skill breakdown, targeted training, and competency assessment
Three transformations redefining AI education: From MOOC to smart courses
Shixizhi 2.0 addresses the three fundamental questions of what to teach, how to teach, and how to demonstrate learning. It does this through three transformations (see Figure 2), distilled into a practical framework: one learner, one competency map; one course, one AI agent; one role, one capability profile.
Figure 2: Shixizhi 2.0's three transformations
Redesigning the curriculum: What to teach. The platform uses industry job data to keep AI-native courses up to date, phasing out outdated content and focusing on emerging core skills such as AI tool application, human-AI collaborative development, intelligent system operations, and AI ethics and compliance. "One learner, one competency map" means that each learner has a personal skill map that stays aligned with their target role in real time. What they learn is determined by both industry demand and their own skill gaps, ensuring that what they study is what they actually need.
Redesigning the learning experience: How to teach. Learning shifts from passive to active, and content formats evolve from traditional MOOCs to smart courses. "One course, one AI agent" means that each course is powered by multiple AI agents working in concert. The multi-agent interactive classroom integrates AI content creation, AI instructors, AI teaching assistants, AI-led discussions, and AI Q&A, while AI coaches provide round-the-clock tutoring support. For fragmented scenarios such as business travel, commuting, and on-site client visits, courses are broken down into multimodal knowledge units — audio, text, infographics, knowledge flash cards — delivered in lightweight, short video formats that can be consumed anytime, anywhere. This shifts learning from person-to-person to network-based dissemination. Powered by AI algorithms, the platform builds foundational knowledge for beginners, maps the shortest transition path for career-switchers, and pushes cutting-edge topics for senior professionals, enabling personalized, tiered learning for every user.
Redesigning assessment: How to prove learning. The platform provides a process-oriented, context-specific, outcome-driven AI capability assessment model. AI-generated questions, AI-led defenses, AI proctoring, and specific hands-on evaluations are offered throughout the learning journey. Process data tracks learning trajectories, and project outcomes replace traditional written exams as demonstrations of competence. "One role, one capability profile" means that assessment results are captured in a dynamic profile that stays aligned with job requirements. Managers can see the distribution of team capabilities at a glance, and individuals have a clear view of their growth trajectory. This means recruitment and promotion have a more reliable basis than with certificates alone.
Value in action: From isolated efficiency gains to an industry talent backbone
Shixizhi 2.0 embeds AI capabilities across the entire learning journey by thoroughly integrating with business scenarios. It delivers tangible results across three fronts: customer transformation, partner enablement, and the university ecosystem.
Huawei's training and certification for carrier customers: An AI core talent matrix. The platform provides comprehensive AI role-based courses and Huawei certifications such as HCIA-AI across four talent domains: computing power, algorithms, data, and AI applications. With the Shixizhi 2.0 Learning Planning Assistant, capability assessments accurately identify skill gaps and recommend personalized learning paths, shortening job planning from two days to just one hour.
Huawei's GTS O3 Community enablement for partner service engineers: End-to-end workflow and bite-sized learning modules. In training and enablement for partners, the O3 Community offers online courses (both advanced and context-based) and regularly runs special programs such as skills competitions and delivery workshops. However, traditional self-paced online learning lacks a complete view of the knowledge system, offers little integration of theory and practice, and makes it hard to stay engaged, due to ineffective formats like long videos. Partners want to learn in short, flexible bursts.
To address this, the community has used the Shixizhi 2.0 Super AI Mentor (see Figure 3) to enable an end-to-end learning process: assess, learn, practice, test, evaluate, write, distill, recommend, and Q&A. At the front end, learners can set goals, get resource recommendations, and produce course content in one place. In the middle, an AI mentor provides one-on-one guidance and context-specific practice through tools like AI drills, memory cards, knowledge graphs, AI Q&A, and AI summaries, enabling short, intensive learning anytime, anywhere. At the back end, online defenses and capability assessments close the loop. This model has more than doubled learning efficiency for Ascend AI courses.
Figure 3: Shixizhi 2.0, an AI- and data-driven digital knowledge learning platform
The O3 Community integrates online courses, interactive Q&A, community activities, and online labs to help partner engineers around the world solve challenges they encounter in their work.
Huawei ICT Competition for the university ecosystem: Large-scale global talent selection. The Huawei ICT Competition has been recognized as a key partner flagship project of UNESCO's Global Skills Academy. Shixizhi 2.0 provides professional exam services for the competition. It can support over 100,000 concurrent test-takers, with assessments that include HarmonyOS ArkTS and Cangjie programming. The competition enables professional selection for university students across more than 100 countries, helping to advance both theoretical knowledge and practical skills in cloud computing, networking, AI, and software technologies.
To date, Shixizhi has served more than 200 tenants, including Huawei Talent, Huawei Certification, Huawei ICT Academy, Huawei Career, O3 Community, Knowledge Network, MA Academy, and the Brave New World program for new employees. It has enabled efficient, self-directed learning for over 4.35 million learners. The first 19 tenants have been upgraded to AI Academies, and 42,000 courses have been upgraded to smart courses. Feedback from the field reflects the value of these courses in the real world: "Course summaries and abstracts are precisely distilled"; "AI Q&A can accurately pinpoint relevant knowledge points, deepening understanding"; "The AI certification assistant has significantly improved efficiency."
The deeper value, however, lies in systematically addressing three persistent challenges in the ICT industry: skills gaps, difficult transitions, and supply-demand mismatches. For individuals, learning is no longer a months-long career interruption but a continual upgrade process embedded in the workflow, with clear pathways for transition and a roadmap for advancement. For enterprises, existing technical talent can be rapidly reskilled, preventing the dual losses of unfilled new roles and underutilized experienced employees, while significantly shortening the return cycle on talent investment. For the industry as a whole, the closed-loop system continually delivers multi-skilled ICT talent who understand both the technical foundations and business scenarios. This strengthens the industry's talent backbone in the AI-native era.
The ceiling for AI is determined by computing power, but the boundaries of the industry are determined by talent. When AI serves as both the subject and the engine of learning, the ideal of education — scalable, high-quality, and personalized — is becoming a reality. Huawei Shixizhi 2.0 invites partners across the industry to join us in helping every ICT professional keep pace with rapidly evolving technology, as we write the next chapter of AI-powered education together.
- Tags: