A society of human-machine co-evolution is approaching. Photo: TUCHONG
The innovation and application of a new generation of artificial intelligence (AI) technologies are reshaping the discipline of sociology. From the founding of sociology up until the widespread deployment of generative AI, human society has been its primary object of inquiry. Now, with machines capable of speaking human language and engaging in interactions, they have become parallel actors alongside humans in the social world. Although machines have not yet assumed social responsibilities, they already function as non-subjective social actors that exert influence on human behavior. As a result, sociology has begun shifting from an exclusive focus on human society to the study of human-machine society, prompting fundamental changes in its objects, subjects, and methods.
Object: from human society to a society of human-machine co-evolution
Regardless of the twists and turns in its development, once AI acquired the ability to interact with humans, it became a non-subjective social actor that humans must confront. At that point, society ceased to be solely human; it began to develop into a co-evolutionary human-machine society. As many scholars have noted, interactions between humans and machines are giving rise to a social reality that is far more complex than that of human society alone. Sociology has two possible strategies in response to this shift.
The first is to ignore the social actions of machines and their consequences. This was once feasible—during a time when machines lacked autonomous social capacities or when their actions had clearly traceable consequences rooted in human decisions. In fact, machines have long existed alongside humans, and sociology traditionally operated under this assumption. However, the turning point brought by generative AI is that machines now possess not only decision-making and action capabilities, but also autonomy. They generate direct, socially significant consequences—such as intelligent capture, analysis, and automated decisions across various transportation systems.
The second strategy is to incorporate the social actions of machines into the scope of sociological research. If these actions and their consequences can no longer be ignored, should machines be treated as subjective actors? While interactions between humans and machines may fall under patterns of competition, coordination, cooperation, or dissemination—and many scholars stress the quasi-actor status of machines—humans and machines are not equivalent social actors. Their interaction is one between a subjective agent and a non-subjective one; humans continue to exercise agency. Machines may exhibit autonomy, but they do not, in essence, possess agency. Technically, it is possible to separate the social actions of machines and their consequences, but in practical terms, the effort is both difficult and costly.
A more feasible approach is to incorporate machines’ social actions and their consequences into the research scope of sociology and treat them as objects of sociological inquiry. These actions act upon human agency, helping to construct a society in which humans and machines co-evolve—with humans as subjects. Sociology thus incorporates machines into its purview not as coequal objects, but as non-subjective actors in a society where humans remain the primary subjects. Accordingly, sociology’s object shifts from human society to a society of human-machine co-evolution.
Subject: from the succession of human society to that of human agency
Sociology was originally inspired by modern physics. Its early aim was to clarify the mechanisms by which society, viewed as a kind of machine, operates—that is, to uncover the principles underlying social functioning. Yet classical sociologists such as Karl Marx, Max Weber, and émile Durkheim transformed Auguste Comte’s principle-oriented project into a value-oriented one. They argued that understanding social mechanisms should serve not only for seeking social principles but also for preserving and developing the human society. This turn toward values gave rise to sociology’s multilayered thematic structure, encompassing topics like social structure, solidarity, and culture. Even today, explorations of inequality, social stratification and mobility, or cultural values—whether descriptive or explanatory—remain aimed at maintaining and optimizing social order. Ultimately, these varied themes converge on a shared concern: social development and succession.
The inclusion of machines in the sociological field has sparked widespread debate—from concerns over machines replacing human labor to fears of human subjugation by machines. This has generated confusion about human survival and agency in a human-machine society, as well as anxiety over the value and meaning of human existence and the potential for social disorder. Amid this fog of opinions lies sociology’s most pressing subject in the era of human-machine society.
First, human agency should be stressed. In human society, human agency is a philosophical subject. However, in a human-machine society, human agency sinks to the first subject of sociology. Whether machines will have agency in the future or not, human agency remains the foundation of human existential value. Without human agency, the maintenance and continuation of humanity lose their significance. In the wave of technological accelerationism, human agency is not a self-evident presupposition but a theme that requires sociological assertion, elaboration, and clarification.
Second, attention must be paid to the development and succession of human-machine society. While classical sociological subjects remain relevant, the object of inquiry has changed. The new focus is a symbiotic human-machine society, grounded in the preservation of human agency. Even if machines are not fully quasi-social actors, they already act as non-subjective ones. Hence, the exploration of social development and continuity must be reconfigured within the human-machine context. While the pursuit of social principles still matters, it is even more critical to uphold sociology’s value orientation: sustaining the development and continuity of human-machine society and, thereby, human agency.
Third, the thematic core of sociology in human-machine society remains structural. Given the foundational emphasis on human agency and the goal of sustaining human-machine society, lower-level subjects will become increasingly contextualized—from ethnicity, gender, and age to education, healthcare, and socialization; from labor, income, and consumption to entertainment, sports, and the body. Yet every such subject is now interwoven with human-machine dynamics. They must account for the influence of machines on human social actions, the ways in which humans train and use machines, and the forms, mechanisms, and consequences of their interaction—that is, the logic of human-machine symbiosis.
Method: from thought experiments to human-machine symbiosis
Comte envisioned sociology as a discipline modeled after physics. Modern physics itself was shaped by the diffusion of measurement techniques and the rise of experimental methods. Most disciplines that study the specific properties of objects have been influenced by such technologies.
Yet sociology did not follow Comte’s blueprint to become a methodologically unified, measurement-driven field. Instead, it evolved into a discipline grounded in specific social contexts, shaped by the knowledge base and methodological capacities of sociologists, and oriented toward addressing real-world social issues. While measurement technology has influenced sociological practice, it has never consolidated sociology’s methodological foundations. Over the past two centuries, sociological methods have remained pluralistic, encompassing a rich blend of thought experiments, meaning interpretation, and empirical analysis.
Dialectic inquiry can be understood as sociology’s version of the thought experiment. From Comte to Anthony Giddens, sociologists focused on constructing knowledge systems have employed dialectics as a central method. Even today, sociological research concerned with macro-level social facts and the broad characteristics of social phenomena continues to rely on this mode of thinking.
Interpretation can be understood as the sociological extension of social philosophy. The constructed nature of society makes the exploration of the value and meaning of social phenomena an important path to understanding society. The pursuit of meaning, in particular, has greatly expanded the imaginative space in which social facts are understood and remains a key tool in guiding public understanding of society. In the context of human-machine society, interpretive methods may yield even greater social impact.
Empirical analysis can be understood as sociology’s application of measurement technologies. Practice here involves not only the use of established principles but also the continuous innovation of methods and techniques. Originating in the French Annales School and further developed in the context of American society, empirical sociology has steadily advanced its tools across multiple stages—data collection, hypothesis testing, and inference—ranging from early analyses of central tendency and dispersion to structural analysis, time-series analysis, network analysis, and simulation. Sociology has never ceased its efforts to approximate complex social phenomena using limited technical means. However, because data capable of capturing the richness of social phenomena remain scarce, and because the application of such methods requires substantial knowledge and skill, quantitative techniques—despite their evolution—have increasingly become a specialized endeavor accessible to only a technically adept few. Even so, the pursuit of generalizable social laws that can be tested using available data remains a core methodological aspiration. In the era of big data, where we mine relational patterns among social entities, quantitative techniques remain indispensable.
It is worth noting that sociological methodology has never operated on an either-or basis. Mixed methods have long been the norm. For example, researchers may use dialectic methods to formulate questions or hypotheses, draw on case studies, surveys, experiments, or big data for evidence, and then use statistical tests or interpretive approaches to explain relational patterns or social meanings. In the study of human-machine society, these diverse methodological traditions retain wide applicability and knowledge-generating power.
However, the advent of AI introduces a new possibility: dialogue with machines, or human-machine methodological co-evolution. This notion of “dialogue” refers not only to machines’ ability to assist in applying existing methods, but also to the potential for collaboratively exploring new, more suitable methodologies. We are not dealing with a sociology-specific machine, but one that has assimilated nearly the full range of human knowledge. Its capabilities can serve as a resource pool for the development of innovative methods. This has already been amply demonstrated in fields such as mathematics, structural chemistry, materials science, and pharmacology.
In short, with the innovation, development, application, and proliferation of a new generation of AI technologies, human society is inevitably evolving into a human-machine society. For sociology, this represents both challenges and opportunities. A historically responsive sociological posture would be to include machines as objects of study, assert human agency as foundational, treat the succession of human agency in a human-machine society as the central subject, and adopt machines as methods.
Qiu Zeqi is a professor from the Department of Sociology at Peking University.
Edited by ZHAO YUAN