What Do Platform Workers Think?: Analysis of Truck Drivers’ Perspectives on Digital Platforms Using GPT-4

Abstract

This study examines platform workers’ experiences and perspectives about digital platforms in the Chinese freight transportation industry based on a large dataset of self-employed truck drivers’ online discussions. Utilizing the state-of-the-art large language model (LLM), GPT-4 by OpenAI, this article creates a novel topic generation framework, Sequential Abstraction, to analyze the topics of concern among truck drivers. Additionally, this study employs GPT-4 to perform sentiment analysis and topic classification tasks, which achieve accuracy rates of 78.13% and 79.17%, respectively, with manually labeled results as benchmarks. The research results highlight the significant economic insecurities Chinese truck drivers face when finding gigs on digital platforms. This study contributes to existing platform work literature by examining workers’ real-time, spontaneous voices surrounding digital platforms. This study advocates for further uncovering the potential of LLMs in facilitating descriptive and explanatory social science research.