The rapid advancement of AI is profoundly reshaping the underlying logic of human knowledge production. From research innovation to cultural communication, a new paradigm of knowledge production marked by human–machine collaboration is gradually taking form. Art criticism now faces an unprecedented challenge: When ChatGPT can generate exhibition reviews in seconds and computer vision can quantify the curvature of the Mona Lisa’s smile, have the theoretical foundations, practical methods, and core values on which traditional art criticism rests begun to loosen? This question forms the starting point of this inquiry.

Impacting traditional paradigm

Over thousands of years of cultural accumulation, traditional art criticism in both China and the West has developed a relatively stable structure centered on human subjectivity, with written texts as its vehicle. Its internal logic and methodological systems have remained comparatively stable, guiding critical practice throughout the history of art. The advent of AI technology, however, has unsettled this stability and created powerful external pressure.

In the Chinese context, AI is shaking this foundation in two ways at once. Through data-based correlation analysis, algorithms reconstruct artistic patterns, challenging philosophical assumptions and historical interpretations. Generative AI blurs the boundaries of originality and authorial intent, unsettling the foundations of the author-centric approach. At the same time, AI’s data-driven and precision-oriented logic quantifies concepts such as qiyun (rhythmic vitality) and deconstructs yijing (artistic conception). It can imitate the outward forms of brushwork, yet it loses the inner vitality and spirit that animate traditional art. As a result, the core assumptions of traditional criticism have become increasingly difficult to sustain.

The rise of AI criticism, grounded in the technical logic of data quantification, model recognition, and probabilistic prediction, can generate exhibition reviews almost instantly and quantitatively analyze stylistic correlations among artworks. It therefore challenges traditional methodologies rooted in philosophical reflection, close textual reading, intuitive perception, and metaphorical symbolism. Clearly, an interpretive path centered solely on human agency can no longer fully encompass the vast range of AI-generated and human–AI co-created artworks.

AI introduces technical metrics that reduce rich aesthetic experience to data indicators such as color contrast and compositional balance. This flattens aesthetic value and calls into question value judgments grounded in human subjectivity. Art criticism is then no longer a spiritual bridge linking artwork, creator, and audience. Instead, it has become a tool-driven operation governed by technological logic, unable to answer humanity’s deeper spiritual need for art or to establish a meaningful value framework in an age of AI-generated content.

Holistic restructuring

AI is democratizing artistic knowledge, enabling ordinary audiences to understand artworks through technological tools. In this process, AI acts as a data-assisted partner, and the subject of criticism is no longer a single group. Instead, a collaborative structure has emerged, comprising professional critics, AI tools, and general audiences. The critic’s core value has shifted from monopolizing knowledge to coordinating human-machine relationships, positioning them as a bridge between the humanities and technology.

Traditional art criticism has long taken human-created artworks as its object, whose core feature is the sensuous manifestation of the human spirit. The rise of AI-generated art, however, has expanded the scope of criticism beyond purely human creations, introducing features of “weak human intent.” Human artistic creation is also increasingly incorporating AI assistance, turning the object of criticism into a composite form of “human creativity + AI implementation.” This shift requires criticism to move beyond the single path of interpreting human intention.

Traditional critical practice is a purely humanistic form of experiential interpretation, while AI criticism is grounded in the data-driven logic of technology. Both approaches have their respective limitations. As digital humanities and art criticism continue to converge, the methodology of critique is undergoing a profound shift from single-subject judgment toward a human–machine collaborative model.

Traditional text-based media have shaped a paradigm in art criticism that privileges logic over intuition and is characterized by linear narrative structure. Art criticism typically proceeds from the artist’s biography to analysis of the work, and then to value judgment. Virtual reality (VR) technology and dynamic visualization have expanded the range of expressive possibilities, making it possible to concretely convey aesthetic experiences—such as spatial rhythm and the emotional force of color—that are difficult for written language alone to express. Criticism thus moves from simply reading text toward an immersive sensory experience. Driven by human–machine collaboration, multimodal interaction, and composite creation, art criticism is no longer a closed textual interpretation but an open field of cognitive practice.

‘Collaborative paradigm’of human-machine symbiosis

This transformation has not been smooth, as it faces both ethical and aesthetic challenges. The ethical challenges center on issues such as copyright attribution for AI-generated content and the alienation of aesthetic sensibilities, while the aesthetic challenges appear in the erosion of artistic aura by technological logic. The way forward lies in a threefold coordination of regulation, data, and capability.

Ethically, addressing algorithmic bias and ambiguous ownership requires the establishment of guidelines for ethical human–machine collaboration. It should be made clear that AI remains a tool, while humans retain ultimate authority over value judgment. In copyright disputes, the principle of human creative intervention should apply: Purely AI-generated content should not enjoy copyright protection, whereas content that reflects a user’s original contributions should be protected, thereby clarifying the boundaries of rights and responsibilities.

Aesthetically, countering value flattening and the dissolution of subjectivity does not require rejecting AI. The key is to reinforce the irreplaceable role of human critics.

A sustainable transformation also requires a clear delineation of roles between humans and machines. AI should focus on efficiency-oriented tasks such as data processing and feature extraction, while human critics should concentrate on creative endeavors such as value judgment and cultural interpretation.

At present, AI is driving a shift in art criticism from the formerly human-centered authoritative paradigm toward a collaborative paradigm of human–machine symbiosis. This is not a simple upgrade of tools, but a systemic transformation that touches the very foundations of the discipline. Future art criticism will ultimately take a new form in which humans and machines each leverage their respective strengths, coexisting and flourishing together. It will preserve the core of the humanistic spirit while resonating with the technological age.

Zhao Chonghua is a professor and director of the Wang Guangqi Research Center at Sichuan Conservatory of Music.