The intersection of digital education and interdisciplinary integration is accelerating, with the "15th Five-Year Plan for Educational Development" explicitly calling for deeper advancement of new liberal arts construction. This initiative centers on breaking traditional academic boundaries to address national strategic needs and contemporary challenges, with the fusion of humanities and sciences serving as a critical pathway. Yet for years, meaningful academic dialogue between the liberal arts and sciences has remained elusive.
The root cause is not methodological variation but epistemological divergence. Humanities research is goal- and value-driven, relying on textual interpretation, historical analysis, and theoretical reflection. Scientific research, by contrast, is rooted in observation and experimentation, transforming subjects into replicable data sets and mathematical models that generate new insights through quantitative reasoning. How can these two realms be bridged? The rapid advancement of generative artificial intelligence offers a transformative solution—recasting the methodological fragmentation arising from differing epistemologies as a problem of knowledge production itself.
In other words, as AI technology iterates, the challenge facing new liberal arts has evolved into a question of how the humanities can leverage AI to construct a novel knowledge ecosystem in which multiple disciplines and actors collaboratively produce knowledge. To address this, we must systematically examine three essential elements of the knowledge ecosystem—production models, structural configurations, and knowledge agents—to understand how AI is reshaping the entire landscape.
AI is transforming knowledge production models in the new liberal arts
A knowledge production model refers to the fundamental mechanism that sustains the ongoing operation of a knowledge ecosystem. Traditional theory on knowledge production transitions describes a shift from Mode 1—purely academic and discipline-specific—to Mode 2, which is application-oriented, involves heterogeneous participants, and is cross-disciplinary. Generative AI has not only accelerated this transition but is pushing knowledge production toward an interactive, generative model. This emerging paradigm is problem-driven and human-computer collaborative, where users continually query the system, prompting iterative refinement of model parameters and generation strategies to yield increasingly tailored solutions.
This new production model creates preconditions for multi-disciplinary and multi-actor collaboration, with AI serving as the operational platform. For instance, humanities researchers define objectives and significance around specific applied problems, asking AI to analyze and interpret those goals to derive methods and tools. Scientists then design algorithms and build models based on the proposed solutions, using AI simulations to test feasibility and refine approaches. This collaborative pattern has already been validated in humanities research. Peking University's Digital Humanities Research Center developed the "Ancient Text Provenance Analysis Platform," establishing a model of science serving humanities scholarship. Similarly, Washington University in St. Louis's AI Humanities Lab extracted stylistic features of 19th-century literary figures from GPT models and designed classification algorithms, producing a detection system capable of distinguishing AI-generated text from genuine human authorship.
Given generative AI's vast information repositories and linguistic capabilities, this new production model allows knowledge agents across disciplines to concentrate on problem-solving, gradually fostering identification with a novel knowledge structure through collaborative engagement.
AI is reshaping the structural configuration of new liberal arts knowledge
Knowledge structure denotes the organizational form and logical interconnections of knowledge. Under the impetus of generative AI, new knowledge production models are driving a reconfiguration of new liberal arts structure, shifting from single-agent dominance to multi-actor interactive networks. Actor-network theory posits that knowledge production is never accomplished by humans alone but emerges from networks involving myriad factors. In such networks, humans are not the only "actors"—technical tools, data, and devices also qualify. Each actor's influence varies in strength. Compared to instruments like microscopes or data equipment, which serve as auxiliary knowledge production tools, generative AI resembles a quasi-human "strong actor."
Consequently, humans no longer occupy an absolute central position in the knowledge structure. Humanities researchers, natural scientists, and generative AI all function as "strong actors" constituting the network, interacting to form an integrated whole. The humanities researcher acts as conductor, grasping the core and underlying logic of the knowledge structure and charting strategic direction. Generative AI plays the role of archivist, handling analysis, organization, and interpretation of knowledge materials. The researcher in the natural sciences assumes the architect's responsibility, designing and assembling the knowledge framework. The interactive generative model both gives rise to this new structure and sustains it through continuous, dynamic interplay among these three actors.
The emergence of multi-actor interactive structures manifests concretely in the reorganization of traditional disciplinary units into a transdisciplinary configuration characterized by problem-orientation, borderless disciplines, and multi-stakeholder participation. This "transdisciplinarity" represents the practical embodiment of the new knowledge structure and the ideal direction for new liberal arts development, with preliminary explorations already underway. Stanford University's CRAFT AI classroom redesign project and AI4ALL initiative, École Polytechnique's Master's program in Visual and Creative AI, and China's 2025 inclusion of "AI Education" as a formal undergraduate major—pioneered by Beijing Normal University—all point toward this trajectory. These efforts are collectively fostering the emergence of a new knowledge subject.
AI is driving the formation of new knowledge agents in the new liberal arts
A knowledge agent refers to both the inheritor and creator of knowledge. In the AI era, knowledge agents within new liberal arts are simultaneously creators driving novel production models and structures, researchers and educators, and learners who internalize and advance these new paradigms. The cultivation of the learner—particularly the latter—serves as a critical indicator of whether a genuinely new knowledge ecosystem has materialized. Scholars in "post-human" studies argue that with advancing information technology, individuals are no longer static entities with fixed knowledge systems and material forms, but dynamic existences shaped through continuous interaction with technology and information.
From this perspective, new liberal arts education must cultivate not merely "knowledge holders" but "dynamic subjects" capable of perpetual learning and knowledge generation in symbiosis with technology. This necessitates preparation across three dimensions: talent development models, curriculum structures, and institutional operations.
Regarding talent development, new liberal arts education must evolve from traditional knowledge transmission toward cognitive generation within interactive frameworks. Conventional humanities education focuses on knowledge accumulation, internalization, and skill development. In the AI era, the educational subject is an embodied sensing agent, making the teaching-learning process a continuous interaction between this subject, technology, and environment. Educators recreate the original context of cognitive emergence, guiding learners to mobilize embodied experiences in discovering real social problems and industrial needs, posing questions to generative AI, and seeking feasible solutions.
At the curriculum level, new liberal arts education will shift from conventional disciplinary divisions to transdisciplinary configurations that connect diverse knowledge agents. Transdisciplinarity entails creating open learning spaces rather than imposing knowledge production norms—expressed concretely through interdisciplinary courses and projects at humanities-technology intersections. Interdisciplinary courses provide learners resources beyond their disciplinary boundaries, enabling problem-guided capability development. Interdisciplinary projects require task-driven engagement, where learners proactively collaborate with multiple actors, including generative AI, and in that process reconstruct themselves as new knowledge agents within the evolving structure.
At the institutional level, new liberal arts operations must dismantle entrenched compartmentalization and move toward open, collaborative networked functioning. Traditional universities manage via departmental silos, whereas a knowledge ecosystem constitutes an open, dynamic network. This structure requires opening academic boundaries to research, industry, and society—making scholarly production and governance responsive to societal and corporate input—while advancing industry-education integration and science-education convergence. Centered on knowledge interaction, generation, and application, it promotes UGSS (University-Government-School-Society) four-dimensional collaborative educational mechanisms, ultimately building a "research-education-industry" synergistic ecosystem.
In sum, within the AI era, new liberal arts signifies both the dissolution of "liberal arts" as a disciplinary boundary concept and its presence as an interactive generative knowledge model, a multi-actor interactive structure, and an ecosystem fostering dynamic subjects. AI's natural language interfaces and algorithmic logic cannot replace human historical experience and intuitive judgment. Therefore, new liberal arts researchers and educators must not merely serve as counterparts in technological interaction, but should assume roles as technology supervisors and decision-makers regarding its direction. They can thereby sustain new liberal arts' foundational advantages in rooting within national cultural traditions and preserving humanistic values, while leveraging its positive role in cultivating mature logical reasoning and well-rounded personalities. Moreover, they can respond to contemporary developmental needs, advancing the optimization and reconstruction of knowledge production models and structures—ensuring societal progress and technological advancement ultimately serve the construction of new knowledge ecosystems, the establishment of novel knowledge agents, and their continuous cultivation, refinement, and advancement.
Comments