A systematic review of conversational artificial intelligence for smoking cessation
More details
Hide details
The University of Queensland, Thoracic Research Centre, Faculty of Medicine, Australia
The Prince Charles Hospital, Queensland Health
The University of Queensland, School of Public Health, Faculty of Medicine, Australia
The University of Queensland, NHMRC Centre of Research Excellence on Achieving the Tobacco Endgame, School of Public Health, Faculty of Medicine, Australia
The Australian e-Health Research Centre, Commonwealth Scientific and Industrial Research Organisation
Publication date: 2023-04-26
Popul. Med. 2023;5(Supplement):A595
Conversational artificial intelligence (AI) (chatbots, dialogue systems and virtual agents) is an emerging tool for tobacco smoking cessation which has the potential to emulate personalised human support and increase engagement. We aimed to determine the effect of conversational AI interventions with or without standard smoking cessation interventions on smoking cessation outcomes among adults who smoke, compared to no intervention, placebo intervention or an active comparator.

A comprehensive search of six databases was completed in June 2022. Eligible studies included randomised controlled trials (RCTs) published since 2005. The primary outcome was sustained tobacco abstinence, self-reported and/or biochemically validated, of at least 6 months. Secondary outcomes included point-prevalence abstinence, sustained abstinence of less than 6 months. Data extraction of cessation outcomes and risk of bias assessments was completed independently by two authors.

Five RCTs met inclusion criteria (n=58,796) from 819 studies; all differing in setting, methodology, intervention, participants and endpoints. Conversational AI interventions included chatbots embedded in multi- and single component smartphone apps (n=3), a social media-based (n=1) chatbot, and an internet-based avatar (n=1). Random effects meta-analysis found participants in the conversational AI enhanced intervention were significantly more likely to quit smoking at the end of the trial compared to control group participants (RR = 1.61, 95% CI (1.21, 2.13), p = .001). High heterogeneity was found between studies (Q(4) = 53.84, p<.001), as well as high overall risk of bias. Loss to follow-up was generally high.

There is limited but promising evidence on the effectiveness of conversational AI interventions for smoking cessation. While all studies found benefit from conversational AI interventions, substantial heterogeneity means the results should be interpreted with caution. Given the rapid evolution and potential of AI interventions, further well-designed RCTs are warranted in this promising area.

Journals System - logo
Scroll to top