Using an Artificial Intelligence Based Chatbot to Provide Parent Training: Results from a Feasibility Study

Abstract

Online parenting training programs have shown to be effective. However, no studies on parent training programs delivered through chatbots have been reported yet. Aim. This study aims to assess the feasibility of delivering parenting skills through a chatbot. Methods. A sample of 33 parents completed a pilot feasibility study. Engagement, knowledge, net-promoters score and qualitative responses were analyzed. Results. A total of 78.8% of the sample completed the intervention. On average, participants remembered 3.7 skills out of the 5 presented and reported that they would recommend the chatbot to other parents (net promoter score was 7.44; SD = 2.31 out of 10). Overall, parents sent a mean of 54.24 (SD = 13.5) messages to the chatbot, and the mean number of words per message was 3. Main themes parents discussed with the chatbot included issues regarding their child’s habits, handling disruptive behaviors, interpersonal development, and emotional difficulties. Parents generally commented on the usefulness of the intervention and suggested improvements to the chatbot’s communication style. Conclusions. Overall, users completed the intervention, engaged with the bot, and would recommend the intervention to others. This suggests parenting skills could be delivered via chatbots.

User experience with a parenting chatbot micro intervention

“…In terms of engagement, participants sent an average of 49.8 messages during a one-time 15-minute intervention. The number of messages is similar to what was found in the study of the beta version of the parenting micro intervention (M = 54.24) (33). Interestingly, another study on a Spanish-speaking chatbot (32) also reported a high average number of messages sent (M = 116).…”

“…Interestingly, another study on a Spanish-speaking chatbot (32) also reported a high average number of messages sent (M = 116). Overall, previous Spanish-speaking chatbot studies conducted in Argentina (32,33) along with the current one have shown a higher engagement than what was found in a longer English-speaking chabot study (M = 17.57 messages) (45). It is possible that participants in Argentina are more open to sharing their issues with a chatbot than English-speaking users.…”

“…Compared to the pilot study (33), the completion rate was lower (78% vs. 66.3%), though the present study included more participants and had a more diverse recruitment effort (online vs. snowball recruitment), making the results more generalizable. The current completion rate is higher than the one obtained in another chatbot study conducted in Argentina (40%), although that intervention took place over a longer period of time (8 weeks) (32).…”

“…To our knowledge, there is one study on a parent training intervention delivered through a chatbot ( 33 ). The study examined the feasibility of delivering the beta version of a parenting chatbot micro intervention to teach parents how to praise their children.…”

Editorial Introduction to Technological Approaches for the Treatment of Mental Health in Youth

“…The authors note that in-person protocols for treatment are typically extensive, 10-25 h per week, thus imperative to further investigate the optimal number of hours for video conferencing and how online and in-person elements in treatment can be combined. Another article by Entenberg et al (2021) assessed the feasibility of delivering parenting tips via an artificial intelligence based chatbot accessed through social media. The study found that most parents completed the intervention, engaged with the bot, and would recommend it to others, which is promising in terms of scalable low-cost interventions to support parenting skills.…”