PDF(Pubmed)
Abstract：
Chatbots can effect large-scale behaviour change because they are accessible through social media, flexible, scalable, and gather data automatically. Yet research on the feasibility and effectiveness of chatbot-administered behaviour change interventions is sparse. The effectiveness of established behaviour change interventions when implemented in chatbots is not guaranteed, given the unique human-machine interaction dynamics. We pilot-tested chatbot-based behaviour change through information provision and embedded animations. We evaluated whether the chatbot could increase understanding and intentions to adopt protective behaviours during the pandemic. Fifty-nine culturally and linguistically diverse participants received a compassion intervention, an exponential growth intervention, or no intervention. We measured participants\' COVID-19 testing intentions and measured their staying-home attitudes before and after their chatbot interaction. We found reduced uncertainty about protective behaviours. The exponential growth intervention increased participants\' testing intentions. This study provides preliminary evidence that chatbots can spark behaviour change, with applications in diverse and underrepresented groups.
摘要：
聊天机器人可以影响大规模的行为改变，因为它们可以通过社交媒体访问。灵活,可扩展，并自动收集数据。然而，关于聊天机器人管理的行为改变干预措施的可行性和有效性的研究很少。在聊天机器人中实施既定的行为改变干预措施的有效性得不到保证，鉴于独特的人机交互动力学。我们通过信息提供和嵌入式动画对基于聊天机器人的行为改变进行了试点测试。我们评估了聊天机器人是否可以在大流行期间增加理解和采取保护性行为的意图。59名文化和语言不同的参与者接受了同情干预，指数增长干预，或者不干预。我们测量了参与者的COVID-19测试意图，并测量了他们在聊天机器人互动前后的待在家里的态度。我们发现保护行为的不确定性降低。指数增长干预增加了参与者的测试意图。这项研究提供了初步证据，表明聊天机器人可以引发行为改变，在多元化和代表性不足的群体中应用。
