The information revolution is rapidly permeating various fields of socio-economic development, promoting digital transformation and forming new quality productive forces. It has facilitated and continuously advanced the digital transformation of education, bringing profound changes to the field. Education has entered a new form of development characterized by “maintaining principles while innovating”. On the one hand, digital education brings “innovation” to the existing educational form, enriching and developing modern education; on the other hand, it is “maintaining principles”, safeguarding the fundamentals and grasping the direction for educational development, embodying the dialectical unity of “change” and “unchanging”, inheritance and development, principle and creativity. It presents the commonalities and characteristics of the modern educational system, and gives rise to a new educational form - new quality education. It is a transformation of education to adapt to new quality productive forces, promoting a qualitative leap and improvement in educational services for economic and social development.

This paper begins with the characteristics of AI innovation and the patterns of talent development, clarifying the shift in the underlying logic of cultivating top-notch innovative AI talent and the three critical challenges involved. Through a review and comparison of typical practices in countries such as the United States, the United Kingdom, and France, it then identifies the common approaches adopted by these countries in addressing the challenges above. On this basis, it further reveals the internal mechanism of unconventional cultivation, proposing that the cultivation system be transformed toward hub integration, elastic agility, and ecological synergy, thereby constructing an educational ecosystem that supports the co-evolution of talent development and innovation systems. Grounded in China’s national conditions and institutional strengths, the paper puts forward practical pathways for the independent cultivation of top-notch innovative AI talent, so as to provide a theoretical basis and practical reference for China to accelerate the development of new competitive advantages in AI talent.

The development of scientific and technological talent has long been explained through two competing theoretical perspectives: universalism and particularism. However, the relevant empirical evidence remains fragmented and inconsistent. A three-level meta-analysis of 25 empirical studies conducted globally (k=130, N=33 087) found that universalistic factors had a statistically significant large effect on scientific and technological talent development, whereas particularistic factors had a statistically significant medium effect. Among specific factors, research impact emerged as the strongest predictor; particularistic factors including educational backgrounds, supervisor’s prestige, and academic networks also showed significant positive associations. The moderating effects of gender, age, and country did not reach statistical significance. These findings address the long-standing theoretical debate between universalism and particularism in the development of scientific and technological talent with more robust evidence, while offering practical implications for maintaining a focus on knowledge contributions, improving access to high-quality research resources, and establishing a universal support system for all scientific and technological talents.

Acceleration is a defining feature of modern society. In his theory of social acceleration, Hartmut Rosa classified the various forms of social acceleration and analyzed three engines driving it. Building on this foundation, this paper proposes a theoretical framework of differential structures and, consequently, introduces a new metaphor—the “differential” — to describe the role of university in an accelerating society. This article argues that social structure is not governed by a single, uniform speed; rather, it consists of fixed components, fast and slow sectors, and a “differential” that mediates between them. This “differential” regulates disparities in speed across different sectors, thereby preventing fragmentation within society. In the context of an accelerating society, the university functions precisely as such a crucial “differential”, serving as an intermediary that sustains the connection between fast and slow sectors. To secure its legitimacy within an accelerating society, the university, as a “differential”, must place greater emphasis on the production of knowledge associated with the slow sector and undergo a process of “re-educationalization” and promote public reflection. In doing so, it can provide direction and boundaries for social acceleration, thereby avoiding the slippery slope of disorderly acceleration.

Local high-level universities constitute the mainstay of the higher education system and serve as critical pivots for implementing the strategy of building a strong education nation at the regional level. Operating under structural constraints characterized by relatively weaker platform capacity, financial resources, and institutional provisions, these universities face overlapping dilemmas in overseas talent recruitment, which manifest as macro-level competitive disadvantages, meso-level organizational deficiencies, and micro-level governance lags. Five recruitment models have been widely adopted in practice, including talent-recruiting-talent, activity-based recruitment, project-based recruitment, cooperation-based recruitment, and institution-based recruitment. However, their effectiveness is highly contingent upon external policy windows and individual relational networks, exhibiting a pattern of “diverse without integration, and effectiveness without robustness”. The study reveals that the predicaments faced by local high-level universities in overseas talent recruitment are not simply attributable to resource scarcity, but rather result from the interplay of path dependence arising from resource dependency, incentive misalignment stemming from institutional isomorphism, and insufficient governance capacity. The pathway to breakthrough should transcend the conventional “policy patchwork” approach and shift toward a systemic project centered on strategic focus, resource integration, platform consolidation, institutional optimization, and governance upgrading, so as to construct a new form of overseas talent recruitment for local high-level universities.

Teaching self-efficacy is a micro-level predictor of the implementation effect of the “National Excellence Program” policy, and mentor support can have a significant impact on the teaching self-efficacy of graduate students in the “National Excellence Program”. Based on the input-environment-output theory and using a mixed research method, this study explores the effects of three types of mentor support on the teaching self-efficacy of graduate students in pilot universities of the “National Excellence Program”. The research found that all three types of mentors independently and significantly positively influence students’ teaching self-efficacy, but there is a slight lack of synergy and some conflict; subject mentors provide adequate support, but there is a hidden risk of academic bias; pedagogical mentor support is relatively stable and is key for graduate students in the “National Excellence Program” to overcome teaching difficulties; practical mentor support is less satisfactory, and its effect on the teaching self-efficacy of graduate students in the “National Excellence Program” needs further optimization. In view of this, in order to help graduate students in the “National Excellence Program” develop a passion for teaching, become competent in teaching, and excel in teaching, and to improve their teaching self-efficacy, it is necessary to establish a collaborative governance mechanism for the three-mentor system to facilitate the full play of overall synergy, set up an evaluation and incentive system to correct misalignments in mentors’ educational concepts, and build a collaborative co?education platform to overcome obstacles to integrated mentoring.

In the process of digitalization of higher education, Global South countries are facing both historical opportunities of technological advancement and structural difficulties. Policy discourse has inherent tension and regional heterogeneity. Employing Critical Discourse Analysis and a postcolonial framework, this study examines 47 higher education digitalization policies from 21 Global South countries. Findings reveal in text analysis the discourse themes focus on digital infrastructure and educational equity, independent talent cultivation, digital sovereignty and discourse power. In discourse practice analysis, regional discourse production presents three models: state-led, internationally dependent, and multi-competitive. In social practice analysis, discourse power structure in digital transformation is marked by the tension between discursive dependence and digital sovereignty. It is argued that China can gather Global South support to co-build an inclusive global digital education governance order through discourse reconstruction, platform collaboration, and independent governance.

Since the late 20th century, the UK has encouraged academics to strengthen public engagement with research, promoted the transformation of engaged research in universities, and enhanced their contribution to the sustainable development and common good. From the perspective of knowledge production, engaged research reflects the people-oriented characteristics of knowledge production: it is oriented toward the pursuit of common good, emphasizes organizational heterogeneity, advocates the establishment of long-term dialogue, encourages transdisciplinary research cooperation, and adopts more diversified quality evaluation criteria. The UK has promoted the practice of engaged research by formulating strategic agreements, increasing project funding, reforming external evaluation, and restructuring professional standards. The transformation of engaged research in UK universities provides valuable implications for China to improve the contribution of university research to social development and advance the people-oriented knowledge production.

In the context of “changes unseen in a century”, the underlying architecture and operational mechanisms of Chinese-foreign cooperation in running schools (CFCRS) are undergoing profound transformations. Drawing on data from CFCRS programs between 1995 and 2025, this study introduces institutional logics theory to systematically examine the trajectory and internal dynamics of their structural evolution. Findings reveal that, taking 2020 as a critical juncture, the dominant force within the CFCRS field has shifted from an interest-driven “market logic” to a strategy-led “State logic”. Driven by this shift, the spatial layout of cooperation has expanded toward “the Belt and Road” countries; educational levels exhibit a pattern of “foundational anchoring and high-level prudence”; and disciplinary configurations are realigning with fields of urgent national need. However, the alternation and collision of old and new logics have generated three challenges: the impairment of transnational trust chains caused by geopolitical shocks, a systemic misalignment between macro?strategic demands and micro-level institutional capacities, and regulatory gaps driven by market inertia. To effectively address this changing era, future efforts should reshape a resilient and diversified global cooperation network, advance the stratified optimization of program structures, and construct a digital governance ecosystem.

Based on survey data from 17,838 graduating Master’s and PhD students (Class of 2025) across 46 Chinese universities, our study operationalizes the boundary into two dimensions: behavioral boundary and guidance boundary. Regarding the behavioral boundary, findings reveal a “Red-Yellow-Green” risk zoning of supervisor behaviors; “sending late-night messages” and “forbidding internships” constitute the Red Zone of unacceptable behavior. Furthermore, a high-quality supervisor-student relationship significantly reduces student rejection of Yellow and Green zone behaviors, but cannot override this Red Line. Regarding the guidance boundary, the supervisor’s role manifests as an “academic core, career support, and humanistic care” trinity. When supervisors additionally act as career planners and life mentors, students’ satisfaction, well-being, and employment rates improve significantly. Finally, the optimal dominance model for co-creation evolves dynamically: from student autonomy (early-stage) to joint consultation (mid-stage) and supervisor gatekeeping (late-stage).

The innovation capabilities of engineering PhD students are crucial for promoting national innovation. Based on questionnaire survey data from 1, 307 engineering PhD students, this study constructs a chain mediation model with academic resilience and academic risk-taking willingness as latent mediating variables to explore the relationship between supervisor support and the innovation capabilities of engineering PhD students. The findings indicate that supervisor support is positively correlated with academic resilience and academic risk-taking willingness among engineering PhD students. Supervisor support promotes the innovation capabilities of engineering PhD students through independent and chain-mediated pathways involving academic resilience and academic risk-taking willingness, with a more pronounced effect on innovation outcomes compared to innovation awareness and the innovation process. Further analysis of the different dimensions of supervisor support reveals that academic support, emotional support, and relational support all significantly enhance the innovation capabilities of engineering PhD students. The university should focus on strengthening supervisor support, and continuously improving the innovation supporting ecosystem.

Per-student funding represents the core mechanism of fiscal investment in vocational education. As vocational education enters a new phase of substantive development and refined governance, the uniform funding model can no longer meet the dynamic and differentiated demands of modern industrial development regarding talent cultivation costs and social spillover effects across various disciplines. At the national implementation stage of a differentiated per-student funding system based on major academic categories, comparative analysis of top-level design schemes from representative provinces—including Sichuan, Zhejiang, Hunan, Henan, Fujian, and Yunnan—reveals innovative approaches tailored to local conditions in areas such as disciplinary classification, coefficient variation, and cost accounting. However, routine implementation faces multiple challenges: conflicting data sources between secondary vocational schools and technical colleges, limited financial coordination capacity at grassroots levels, performance evaluations that prioritize form over substance, and structural downward pressure on student enrollment. To deepen this reform, modern fiscal functions must be strengthened as strategic governance tools. This requires implementing categorized full-cost accounting, introducing cross-period smoothing mechanisms for capital expenditures, elevating the focus of fiscal coordination, and establishing a composite funding model that integrates fixed-cost support, differentiated funding across disciplinary categories, performance-based funding, and stability subsidies. Such measures will effectively leverage fiscal funds’ structural regulatory role in supporting the development of a modern industrial ecosystem.