Reproduce “ChatGPT on the Road: Leveraging Large Language Model-Powered In-vehicle Conversational Agents for Safer and More Enjoyable Driving Experience” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yeana Lee Bond, Mungyeong Choe, Baker Kasim Hasan, Arsh Siddiqui, Myounghoon Jeon
Researchers at Virginia Tech developed and evaluated "CARA," a ChatGPT-based in-vehicle conversational agent, finding that its use significantly improved driving stability and was highly preferred by drivers. The agent enabled diverse, multi-turn conversations that went beyond traditional command-based interactions, enhancing perceived competence, animacy, and affective trust.
Key insights
- LLM-powered agents can support more stable driving performance in Level 0 automation contexts, challenging the assumption that all conversational interactions inherently increase driver distraction.
- The quality and adaptiveness of interaction, particularly the ability to engage in user-initiated, human-like conversation, can positively influence cognitive load distribution and driving behavior.
- Anthropomorphic qualities and the ability to engage empathetically significantly influence user perceptions, leading to higher ratings in competence, animacy, and especially affective trust in AI agents.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.