Willingness to utilize autonomous vehicles following accidents: A fresh perspective from mixed-methods research

While autonomous vehicles (AVs) show promise, several challenges remain in their implementation. In this regard, adverse incidents can alter public perceptions and acceptance of this technology. This study used latent Dirichlet allocation (LDA), a random-parameters ordered probit model and a structural causal model to investigate public concerns regarding AV incidents and their influence on the public's willingness to adopt AVs. Using LDA to analyse mass media data, this study identified seven latent variables related to AV incidents. To consider changes in public attitudes following incidents, attitude was introduced as a new latent variable. The impact of various variables on public willingness to use AVs was analysed using a random-parameters ordered probit model. The findings indicated that perceived trust, attitude, perceived risk, mass media, perceived value, brand effect, privacy concerns and policies and regulations are crucial factors influencing AV adoption. Subsequently, a causal structure model was developed to determine the inter-relationships between variables. The model indicated that policy interventions increased public willingness to adopt AVs by 28%, suggesting that policymakers' interventions help create an early market for AVs. In addition, good branding and mass media campaigns influence the public's psychological characteristics and encourage the use of AVs. These findings hold crucial implications for mitigating the negative impacts of adverse incidents, fostering public acceptance and providing valuable insight for theoretical understanding and practical implementation of AV usage.