AI-driven social science research leverages the deep integration of AI technologies as a fulcrum to advance the systematic transformation of social science research. Image generated by AI

AI-driven social science research leverages the deep integration of AI technologies as a fulcrum to advance the systematic transformation of social science research across multiple stages—including question identification and formulation, theoretical framework construction, data collection and analysis, and the presentation of findings—thereby fostering a new research paradigm characterized by human–machine collaboration. Underpinned by the combined impetus of AI and the humanities, this emerging paradigm integrates algorithmic analysis with social inquiry and promotes the co-production of research outcomes through sustained human–machine collaboration.

However, AI-driven social science research also entails multiple risks. One concern is that it may not readily foster original research imbued with human warmth. The social sciences, which seek to explore human society and its developmental patterns, are human-centered intellectual endeavors in which meaning remains central. Current AI systems lack distinctive human capacities—such as personal interests, intuitive perception, imaginative thinking, cultural understanding, humanistic concern, and poetic insight—and do not possess the same depth of critical thinking and imagination as human researchers. These limitations may constrain the production of significant original research grounded in genuine social concerns.

Another concern is the potential weakening of social scientists’ critical capabilities. The distinctiveness of knowledge production in the social sciences lies in the presence of researchers: They enter social contexts through fieldwork, on-site observation, and social experimentation, generating knowledge through the interpretation of social action and meaning. Yet the growing application of AI is reshaping this mode of presence, as “algorithmic presence” risks displacing direct human engagement. Overreliance on AI may encourage intellectual complacency, facilitate the spread of misinformation, diminish creativity, and gradually erode critical thinking and independent judgment.

AI may also undermine the reliability and objectivity of social science research. In recent years, instances of using AI systems to deceive opponents or exaggerate capabilities have become increasingly common. Social scientists may find it difficult to delve beneath the technological surface and fully grasp the internal logic and inferential pathways of algorithms, and they may even be misled without realizing it. In addition, many AI applications exhibit biases related to gender, race, culture, language, and ideology, posing further challenges to the objectivity of social scientific inquiry.

Finally, AI ghostwriting raises concerns about academic credibility. AI systems are now widely involved in drafting papers, monographs, reports, and grant applications, yet their contributions are not always fully disclosed or appropriately acknowledged. Existing academic oversight mechanisms also lack effective means to detect such use. The concealment of AI involvement may therefore give rise to a broader accountability crisis within the academic community.

To mitigate these risks, AI developers, social scientists, and academic evaluators should draw on their respective expertise to build an integrated and innovative application framework.

Guiding the coordinated and mutually beneficial development of the social sciences and AI: Deliberative platforms that recognize deep cultural meaning and accommodate diverse values can help foster closer collaboration between AI designers and social scientists in the development, application, and evaluation of AI systems. At the same time, social science researchers should deepen their understanding of AI itself. Clarifying what AI systems can and cannot do, grasping their operational logic and algorithmic principles, and ensuring that technological tools serve social science research—rather than dictate it—are all essential steps.

Constructing a scholar-centered human–machine collaborative research model: AI should play a broader role in auxiliary technical tasks such as data organization and analysis, as well as pattern identification and model development, while the formulation of research questions, the construction of research frameworks, the interpretation of findings, and the drawing of conclusions should remain under the guidance of human scholars. Effective human–machine collaboration depends on sustained and rigorous interaction, with social scientists determining the rhythm, mode, direction, and scope of engagement. Outputs generated by AI should also be subject to critical scrutiny by human scholars before being incorporated into research conclusions.

Strengthening interaction and exchange between the subjects and objects of academic evaluation in the social sciences: With technological advancement, the objects of social science research are expanding from human-centered networks of social action to a composite relational system of “human–machine–society.” Research approaches are evolving from pure explanation toward an equal emphasis on discovery and explanation, while research methods increasingly exhibit interdisciplinary and hybrid characteristics.

In this context, decentralized technologies such as blockchain can be employed to construct an open and dynamic “field of academic evaluation,” incorporating evaluators, evaluatees, AI agents, and the public into a multidirectional interactive ecosystem. Within this ecosystem, evaluation is no longer a static, one-way judgment but becomes a process of pluralistic social interaction and collaborative innovation. Evaluation results are not intended to provide fixed “correct answers,” but rather to foster a space for open dialogue that accommodates diverse viewpoints. Research outputs, in turn, move beyond fixed forms of representation to become an ever-evolving network of social knowledge.

Xu Dongbo is a research fellow from the Institute for the Development of Socialism with Chinese Characteristics at Southeast University. Zhang Shuangzhi is an associate research fellow from the College of Teachers at Chengdu University.