2023/08/21 by Kuc, Joanna, Oana-Maria Camburu, Lametti, Daniel R. +1
Arts and Humanities · Psychology · #Arts and Humanities #Computer Sciences #Conversational Bots #Digital Humanities #FOS: Psychology #Large Language Models #Medicine and Health Sciences #Mental Health #Physical Sciences and Mathematics #Psychiatry and Psychology #Psychology #Self-Report Measures #Social and Behavioral Sciences #Technology-Assisted Therapy
paper · doi:10.17605/osf.io/3utsm
Traditional journaling, despite its mental health benefits, often lacks structure, immediate feedback, and guidance, impeding emotional pattern recognition and consistent engagement. Emerging digital technologies, particularly large language models, represent a promising avenue for creating structured, interactive tools to improve introspection and self-awareness. Utilising the services provided by OneReach.AI, we developed a conversational bot to encourage daily journaling. We aim to recruit 60 participants, native or fluent English speakers, over 18 years old, through social media, student advertisement channels across different universities and word-of-mouth referrals. For two weeks, the bot will issue reminders and semi-weekly, Generative Pre-trained Transformer 4 (GPT-4) generated summaries of journal entries. Participants will be randomised into three groups; each group will receive a different type of summary: emotional content, cognitive content, or a neutral “thank you for your entries” message. We wish to assess group assignment’s influence on user satisfaction and engagement, evaluate journaling’s mental health benefits and investigate the influence of summary type on the subsequent entries’ sentiment.