Abstract: Artificial intelligence (AI), particularly exemplified by tools such as ChatGPT, has emerged as a key player in the landscape of content co-creation, breaking away from the long-standing reliance on humans as co-creators. However, most existing research has focused on the foundational concepts and practical applications of human-AI content co-creation, with little attention paid to the differences in individuals’ willingness to co-create content with AI compared to humans. To address this gap, the current study adopts cognitive load theory and self-efficacy theory to explore the mechanisms by which individuals’ content co-creation intentions are influenced based on the type of co-creator. Specifically, this research aims to answer two questions: (1) What are the underlying psychological mechanisms by which different co-creators (human vs. AI) influence individuals’ content co-creation intentions? (2) What are the boundary conditions that affect the influence of different co-creators on individuals’ content co-creation intentions?
To explore these questions, this study formulates three research hypotheses, which were tested through three primary studies and an additional supplementary study. In Study 1, a between-subjects design with a single factor (co-creator types: human vs. AI) was utilized to examine the psychological mechanisms through which co-creator types affect content co-creation intentions. The sample consisted of 168 undergraduate students from a comprehensive university in China. Analysis of variance (ANOVA) and bootstrapping for mediation analysis were employed to test hypotheses 1 and 2. In Study 2, the sample and task scenarios were altered and 216 participants were recruited from the Credamo platform to replicate the findings of Study 1. Study 3 employed a 2 (co-creator types: human vs. AI) × 2 (task types: cognitive analytical task vs. emotional social task) between-subjects experimental design to examine the moderating effect of task types on the relationship between co-creator types and content co-creation intentions. A total of 272 participants were recruited from the Credamo platform, and hypotheses 3, 3a, and 3b were tested using ANOVA and bootstrapping for moderating effects. Study 4, a complementary study, recruited 180 participants from the Credamo platform to examine the influence of technical proficiency on the relation between co-creator types and content co-creation intentions.
The research reveals that (1) The effect of co-creator types on content co-creation intention was mediated by cognitive load and creative self-efficacy. Specifically, co-creating with AI decreased individuals’ cognitive load and thereby increased their willingness to co-create content. In contrast, co-creating with humans enhanced individuals’ creative self-efficacy, and thereby increased their willingness to co-create content. (2) The type of task moderated the mediation process. For cognitive analytical content co-creation tasks, co-creating with AI enhanced content co-creation intention, whereas for emotional social tasks, co-creating with humans enhanced content co-creation intention. (3) Technical proficiency served as a boundary condition affecting human-AI content co-creation intention. Individuals with higher technical proficiency experienced a more significant reduction in cognitive load and enhanced creative self-efficacy, which increased their intention to co-create content.
This research contributes in three key ways. (1) While prior studies have examined individual collaboration willingness in human-machine collaboration tasks based on predefined rules and division of labor, this paper focuses on content co-creation tasks that involve deeper human-AI interaction, and further investigates individuals’ willingness to engage in human-AI content co-creation, and thereby enriches the theoretical understanding of human-AI content co-creation. (2) While prior research has primarily focused on content co-creation intentions among human team members, this paper extends the scope by comparing the psychological mechanisms underlying different co-creators’ effects on individual content co-creation intentions, thereby enhancing the theoretical understandings of content co-creation intentions. (3) The study delineates the boundary conditions for the impact of task types on individual content co-creation intentions, thereby expanding the research framework for understanding human-AI content co-creation.

Key words: human-AI co-creation, content co-creation, cognitive load, creative self-efficacy, technical proficiency

摘要： 人工智能（AI）在内容共创中扮演重要角色。然而,相比于与人共创,与AI共创时个体内容共创意愿表现出何种差异?基于认知负荷理论和自我效能理论,通过4项情景实验,考察了不同共创主体对个体内容共创意愿影响的心理机制及作用边界。研究发现：（1）与AI共创有助于降低个体的认知负荷,进而提升内容共创意愿;与人共创有助于增强个体的创造自我效能,进而提升内容共创意愿。（2）任务类型调节了共创主体类型对内容共创意愿的影响,即对于认知分析型内容共创任务,与AI共创有助于提升内容共创意愿,而对于情感社会型内容共创任务,与人共创有助于提升内容共创意愿。

关键词: 人机共创, 内容共创, 认知负荷, 创造自我效能, 技术熟练度