Emerging generative AI divide: Personal, positional, and resource-based factors associated with use and reasons for non-use
Atsushi Nakagomi, Noriyuki Abe, Takahiro Tabuchi · Telematics and Informatics · 2025
• One in five adults in Japan reported using generative AI in early 2025. • Generative AI use was higher among younger individuals, men, and the advantaged. • Personality, psychosocial, and digital factors were linked to AI use. • Distinct patterns of factors associated with each reason of generative AI non-use. • Findings reveal personal, positional, and resource-based barriers in the divide. As generative artificial intelligence (AI) tools become increasingly integrated into everyday life, concerns are growing over unequal adoption and the potential emergence of an “AI divide.” Guided by the Resources and Appropriation Theory (RAT), this study investigated the factors associated with generative AI use and reasons for non-use in a population-based sample of 13,367 adults in Japan. In January 2025, approximately 21.3 % of respondents reported using generative AI in the past 12 months. Modified Poisson regression showed that AI use was more prevalent among younger individuals, men, those with certain personality traits (openness and agreeableness) and those experiencing psychological distress (personal factors). In addition, those who were highly educated, married, childless, living in urban and less socioeconomically deprived areas, and employed in certain occupational categories (positional factors) were more likely to use generative AI. Higher-income individuals, those with strong friend-based social networks, frequent digital engagement (smartphone and social media use), and igher digital health literacy scores (resource-based factors) were more likely to adopt AI. Among non-users, the most commonly cited reason was “not necessary” (39.9 %). We observed patterns in the factors associated with specific reasons for non-use, linked to age, gender, personality traits, education, and digital literacy. For example, younger adults were less likely to report barriers related to skills or security concerns, and instead cited a lack of attractive services. These findings highlight that generative AI adoption is shaped by a wide range of personal, positional, and resource-based. Without inclusive strategies, these disparities risk compounding existing inequalities.