A five-month prospectives study of frequency of generative AI use and workplace loneliness

Abstract

The study employed a 5-month prospective design from February to July 2026. A total of 2,000 participants were recruited through Rakuten Insight, Inc. (https://insight.rakuten.co.jp/), a large internet survey company, from its panel of approximately 2.2 million registered users. Eligibility criteria were being a man or woman aged 20–59 years, employed by a company or organization in Japan, working full-time for at least 35 hours per week. Exclusion criteria was individuals who were on sick leave or maternity leave. Stratified sampling was used to obtain equal numbers of participants by sex and age group: 20–29, 30–39, 40–49, and 50 years or older. Participants completed a self-administered online questionnaire assessing generative AI use, workplace loneliness, and demographic and occupational characteristics. A follow-up online questionnaire survey was conducted at 5 months later, asserting workplace loneliness again. We did not ask questions about workplace loneliness to baseline respondents who retired or changed employer during the follow-up.A total of 1,749 participants (87.5%) responded to the follow-up survey. Forty-five participants who quit or retired from employment were not asked about generative AI use or workplace loneliness. Data at follow-up were available for the remaining 1,704 (85.0%) participants.