Introduction

Recently, General Secretary Xi Jinping sent a letter to all faculty and students of four transportation universities, encouraging them to uphold the educational philosophy of “practical learning and practical work,” inherit and promote the spirit of the Westward Migration, focus on major national strategic needs, strengthen independent technological innovation and talent cultivation, and achieve more breakthroughs in the deep integration of industry, academia, and research. The university has deeply studied and implemented the important spirit of Xi Jinping’s letter, firmly grasping the inherent consistency and mutual support of educational development, technological innovation, and talent cultivation, continuously transforming educational advantages, talent advantages, and innovation advantages into development advantages, competitive advantages, and strategic advantages.

Currently, artificial intelligence is reconstructing human social production and lifestyle at an unprecedented speed, and seizing the high ground of global AI development has become an important support for our country to build international competitive advantages and win great power games. Ultimately, technological competition is a competition for talent and education. In the face of this unprecedented transformation, we recognize that AI education is facing “three structural challenges,” but AI itself provides us with a new way to break barriers and reconstruct paradigms. The “AI + Education Action Plan,” jointly issued by the Ministry of Education and four other departments, clearly requires leveraging AI as an engine for educational transformation and proposes specific requirements such as “building AI learning communities and gathering open-source courses” and “conducting achievement certification to encourage faculty and students to participate in open-source ecosystem construction,” providing direction and deployment for AI education reform. Shanghai Jiao Tong University focuses on cultivating high-quality talent capable of thriving in the intelligent era, seizing opportunities, and consistently using AI as a key lever for enhancing educational capabilities, directly facing challenges, and promoting deep faculty and student participation in AI open-source ecosystem construction, paving a new path that integrates talent cultivation with ecosystem building.

Facing the “Three Constraints” of the AI Era

In the context of rapid iteration of AI technology and continuous upgrading of industrial demand, universities, as the main battlefield for talent cultivation, face multiple challenges in teaching, practice, and resources.

1. Knowledge Barriers: The development of disciplines lags behind technological leaps. Traditional academic systems act as invisible walls, isolating knowledge transfer into fortresses, limiting students to a single disciplinary perspective, making it difficult to form cross-domain innovative thinking. Additionally, the speed of classroom knowledge updates lags far behind the evolution of AI technology, resulting in teaching content often failing to keep pace with the times, leaving students “holding old maps, struggling to find new continents.” This rigid barrier severely restricts the emergence of interdisciplinary innovative talent.

2. Supply-Demand Misalignment: Skills training is disconnected from industrial practice. As AI applications profoundly reshape the labor market, traditional skills reliant on mechanical repetition and rule-based operations face severe challenges of being replaced by “digital employees.” Currently, there is a significant gap between the talent supply from universities and the actual needs of industries: on one hand, various sectors urgently require the implementation of AI scenarios; on the other hand, graduates generally lack real engineering practice experience, making it difficult to quickly translate theoretical knowledge into productivity for solving complex scenarios. This “disconnection between learning and application” not only weakens students’ employment competitiveness but also makes it hard for them to adapt to the rapid iterations of the intelligent era.

3. Resource Scarcity: Innovative exploration is constrained by computing power bottlenecks. Computing power is the “source of motivation” in the intelligent era, and cutting-edge courses heavily rely on AI innovation resources such as computing power, data, models, and tools. However, most universities struggle to bear the enormous investment in intelligent computing clusters and lack the capacity to maintain professional operational teams. Constrained by shortcomings in AI training environments, high-level teaching and research exploration empowered by AI often become a source of water without a stream. The computing power gap has become the biggest bottleneck restricting faculty and students from deeply participating in AI ecosystem construction and producing original results. Without a fertile “research soil,” it is challenging to cultivate innovative fruits that will lead the future.

Reconstructing the Educational Ecosystem with Open Source Spirit

In the face of these challenges, mere slight adjustments to traditional educational models are insufficient to achieve breakthroughs. The open-source orientation clearly defined in the Action Plan provides us with a way to break through—leading with an open-source spirit, breaking barriers, integrating resources, and collaborating on innovation to reconstruct an educational paradigm suitable for the AI era, achieving synchronous resonance between talent cultivation and industrial development.

1. From “Knowledge Transmission” to “Open Source Collaborative Creation”: AI breaks the temporal and spatial barriers to knowledge acquisition, and the open-source spirit makes it possible to innovate while “standing on the shoulders of giants.” AI is not only an object of learning but also the core engine empowering personalized, project-based learning. In the teaching paradigm of “AI + Human Intelligence (AI + HI),” by introducing multi-agent interaction mechanisms and integrating multi-domain expert models, we reshape the learning ecosystem of human-machine collaboration. Through “open-source courses,” we establish a community-based sharing and feedback mechanism, allowing cutting-edge research results, frontline industry practices, and immediate social needs to rapidly translate into teaching content, making learning no longer limited to the classroom or textbooks, and achieving knowledge iteration with “zero time difference” as much as possible.

2. From “Skill Executors” to “Human-Machine Collaborative Innovators”: The core competitiveness of future talent lies not in how much established knowledge they possess but in their ability to harness AI to solve complex engineering problems. We must actively guide students to abandon the anxiety of “being replaced” and focus on cultivating “enhanced innovators” with human-machine collaboration capabilities. We should promote the establishment of a long-term mechanism for deep integration between schools and enterprises, bringing cutting-edge industry practices into the classroom, transforming real pain points from enterprise R&D and applications into practical AI projects for universities, and implementing “real problems with real solutions.” In frontline practical tasks, we hone students’ engineering capabilities, allowing them to “learn to swim in the storm of practical challenges,” and quantify and certify their contributions, forming a lifelong “digital skills passport,” truly realizing the leap from “credential-based” to “capability-based”.

3. From “Resource Islands” to “Inclusive Shared Computing Power Base”: Leveraging national strategic forces and deep industry-education integration, we must seize the historic opportunity to build an efficient collaborative computing power network. We should promote domestic computing power to “enter universities, classrooms, and research,” achieving true “computing power equality” and “educational equity.” Shanghai Jiao Tong University is constructing the “Zhiyuan No. 1” thousand-calorie intelligent computing cluster, promoting large-scale domestic computing power into campus, which not only addresses the shortage of training and inference resources but also lowers usage thresholds and stimulates faculty and student engagement. At the same time, by learning, developing, and innovating in a controllable software and hardware environment, we fundamentally strengthen the security foundation of China’s AI ecosystem, promoting vigorous original innovation.

Building the “Qiwuh Learning Community” as a New Engine for Talent Development

In the face of change, Shanghai Jiao Tong University, based on the concept of reconstructing an open-source educational ecosystem, combines its own educational advantages, transforming theoretical exploration into practical action. By building a specialized talent cultivation platform, we integrate open-source spirit, computing power resources, and industrial needs throughout the talent cultivation process, providing replicable and promotable practical samples for AI education reform.

We are actively planning to collaborate with high-level universities, research institutions, and leading technology companies to create a national AI practical talent cultivation platform—“Qiwuh Learning Community.” “Qiwuh” aims to enlighten wisdom and open the door to innovation; it also seeks to understand laws and internalize engineering qualities. Our core goal is to cultivate high-quality talent for the intelligent era, gathering high-quality open-source courses and introducing advanced domestic computing power to construct a closed-loop of “theory—practice—innovation.”

We will gather thousands of excellent open-source micro-courses, breaking down the “walls” between universities and enterprises, creating an immersive learning environment that integrates theory and practice; introduce domestically produced, controllable large-scale advanced computing power, transforming it into an accessible innovative resource space for frontline faculty and students, solidifying the digital foundation for engineering practice; deepen the “challenge-based” mechanism with leading enterprises, implementing a new model of industry-education integration where “enterprises propose problems, universities lead topics, tackle the same problems together, and jointly evaluate results”; construct a diversified talent evaluation system, establish classified and graded achievement certification standards, and bridge the “last mile” of mutual recognition of achievements between universities; connect quality entrepreneurial resources, empowering students for high-quality employment and cross-border innovation. Let the “Qiwuh Learning Community” truly become a “training ground” for domestic computing power ecosystems and an “accelerator” for the growth of outstanding innovative talents.

AI education is a systematic project that must fully leverage the advantages of the national system while also stimulating market vitality. “One flower alone does not make spring; a hundred flowers in bloom fill the garden with spring.” The essence of open source is connection and symbiosis. We should gather innovative synergy with the open-source spirit, solidify the digital foundation with independent computing power, and jointly compose a new chapter in the high-quality development of AI education in China, allowing every innovative dream to take root and sprout in the fertile soil of open source, injecting continuous innovative momentum into the construction of an educational powerhouse.