Feasibility, Usability, and Promise of a Parent Management Training Using a Generative Artificial Intelligence...

Access to behavioral parenting interventions remains limited due to barriers such as cost, time constraints, and transportation, emphasizing the need for more accessible and reliable means to support caregivers of children with behavioral problems.

Assess the feasibility of a Generative Artificial Intelligence (GAI) platform (ParenteAI.V1) to facilitate and enhance the delivery of Parent Management Training (PMT).

Participants were 22 caregivers of children ages 6-12 years seeking services for disruptive behaviors. The intervention consisted of 8 modules of Parent Management Training (PMT) on the ParenteAI platform through a conversational agent, Pat. A psychology student facilitated the delivery of 4 of the modules during 4 live sessions, where they guided caregivers through the platform and their interactions with Pat. The remaining 4 PMT modules were completed independently by caregivers between the live sessions where they interacted with Pat alone.

Seventeen participants completed the study protocol. Caregivers reported high satisfaction with the intervention, likelihood of recommending it, and a strong therapeutic alliance with Pat. Caregivers exchanged a high number of messages with Pat and completed most PMT modules. There were significant pre- to post-intervention decreases in child externalizing (p < .001, d = 1.00) and internalizing (p = .019, d = 0.63) symptoms, as well as significant decreases in caregiver depression (p = .005, r = 0.68), anxiety (p = .002, r = 0.79), and stress (p = .008, d = 0.74).

Overall, caregiver feasibility, utility, and satisfaction were high. Caregivers reported significant improvements in their children's behavior and in their own mental health. These findings highlight the feasibility, usability, and potential scalability of combining therapists with AI for addressing barriers to traditional parenting programs.