Prevalence of ChatGPT use in a nationally representative Australian community sample

Abstract

Background: ChatGPT and other generative AI tools have dramatically changed the health information environment. Discussions about their benefits and risks need clear evidence about who is using them and why. This study aimed to examine ChatGPT use for health purposes in Australia.   Methods: The sample was recruited from a nationally representative Australian panel in June 2024. Participants completed self-reported survey questions about their use or intended use of ChatGPT for health purposes. Scores were weighted to Australian population benchmarks using propensity scores and associations were assessed using simple logistic regression.   Results: Of the 2,034 participants (50.1% male; 48.4% female) who took part, 9.9% reported that they had asked ChatGPT a health question in the past 6 months (95%CI: 8.5%‚ 11.4%). This was more likely for people born in a non-English speaking country, who speak a non-English language at home, and who have low health literacy (p = <0.05). The most common health questions related to learning about a specific health condition (47.6%), finding out what symptoms mean (37.4%) or what actions to take (35.8%), and understanding medical terms (34.8%). 38.8% of those who knew about ChatGPT but had not yet used it for health purposes would consider asking the platform a health question in the next 6 months. The most common anticipated tasks related to learning about a specific health condition (18.1%), understanding medical terms (16.8%), or finding out what symptoms mean (16.4%)   Conclusions: As usage increases, we need to equip community with knowledge and skills to use these tools safely.