Scholarlink

<p>The present study aimed to develop and validate the Artificial Intelligence Use Motivations Scale (AIUMS)—a multidimensional measure designed to assess the motives and tendencies associated with engagement in generative AI systems (e.g., <i>ChatGPT</i>). Data were collected from 1296 Iranian adults (61.3% female; mean age = 28.74&#xa0;years, SD = 8.61) through an online survey. An exploratory factor analysis (EFA) conducted on half of the total sample (<i>n</i> = 648) initially yielded a 6-factor empirical structure. However, several conceptually distinct but intercorrelated dimensions were merged within those factors. Therefore, based on theoretical frameworks and expert review, a more fine-grained 11-factor model was tested in the confirmatory analysis. The CFA demonstrated excellent fit (<i>χ</i><sup>2</sup>/df = 1.53, RMSEA = .028, CFI = .986, TLI = .979), supporting the theoretical multidimensionality of AI-related motivations. The final model encompassed perceived usefulness, self-efficacy, availability, perceived trust, safety, seeking pleasure, social interaction, escapism, empathy/anthropomorphism, tech-based social prestige, and personal innovativeness, with all subscales showing good reliability (<i>α</i> = .71 − .85). Multi-group CFA confirmed full measurement invariance across gender. Convergent and discriminant analyses showed two motivational pathways: (i) a compensatory–affective system positively associated with problematic ChatGPT use (PCU), social anxiety, and digital addiction; and (ii) a competence–adaptive system associated with higher self-esteem and social support. Hierarchical regression showed that motivational factors explained an additional 15% of variance in PCU beyond emotional, personality, and digital–behavioral predictors. Overall, the AIUMS offers a psychometrically sound framework for understanding motivational mechanisms underlying adaptive and maladaptive engagement with generative AI systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The Artificial Intelligence Use Motivations Scale (AIUMS): Development, Validation, and their Role in Problematic Generative AI Use (PGAU)

  • 期刊论文

  • 发表日期：

    2026年5月30日

查看全文
  • Mehdi Akbari,
  • Shirin Mollaali Farkhani,
  • Mahsa Golmohammadi,
  • Mitra Mirjalili,
  • Fatemeh Yasaghi,
  • Reza Hatampour,
  • Mark D. Griffiths

摘要

The present study aimed to develop and validate the Artificial Intelligence Use Motivations Scale (AIUMS)—a multidimensional measure designed to assess the motives and tendencies associated with engagement in generative AI systems (e.g., ChatGPT). Data were collected from 1296 Iranian adults (61.3% female; mean age = 28.74 years, SD = 8.61) through an online survey. An exploratory factor analysis (EFA) conducted on half of the total sample (n = 648) initially yielded a 6-factor empirical structure. However, several conceptually distinct but intercorrelated dimensions were merged within those factors. Therefore, based on theoretical frameworks and expert review, a more fine-grained 11-factor model was tested in the confirmatory analysis. The CFA demonstrated excellent fit (χ2/df = 1.53, RMSEA = .028, CFI = .986, TLI = .979), supporting the theoretical multidimensionality of AI-related motivations. The final model encompassed perceived usefulness, self-efficacy, availability, perceived trust, safety, seeking pleasure, social interaction, escapism, empathy/anthropomorphism, tech-based social prestige, and personal innovativeness, with all subscales showing good reliability (α = .71 − .85). Multi-group CFA confirmed full measurement invariance across gender. Convergent and discriminant analyses showed two motivational pathways: (i) a compensatory–affective system positively associated with problematic ChatGPT use (PCU), social anxiety, and digital addiction; and (ii) a competence–adaptive system associated with higher self-esteem and social support. Hierarchical regression showed that motivational factors explained an additional 15% of variance in PCU beyond emotional, personality, and digital–behavioral predictors. Overall, the AIUMS offers a psychometrically sound framework for understanding motivational mechanisms underlying adaptive and maladaptive engagement with generative AI systems.