Paper Number
ECIS2026-1548
Paper Type
SP
Abstract
Generative AI (GenAI) is reshaping online labour markets. Existing research mainly examines general-purpose GenAI, such as ChatGPT, and focuses on aggregate outcomes, including falling demand and compressed prices in easily automated tasks, while revealing little about the demand for work skills and the role of platform-embedded GenAI. We explore how platform-embedded GenAI affects work skills. Leveraging logo design job posts before and after the launch of an early-stage platform-embedded logo-AI tool on the online labour market EPWK, we investigate its effect on requested skill diversity using a difference-in-differences design and a new large language model-based skill extraction and embedding framework. We find that logo jobs exhibit higher skill diversity than other design jobs after the platform introduced logo-AI, and that stronger competition among freelancers partially mediates this effect. Our findings indicate that changes in skill demand in online labour markets are an outcome of introducing platform-embedded GenAI.
Recommended Citation
Sun, Fang and Liu, Libo, "Exploring The Effect Of Platform-Embedded Generative Ai On Skill Diversity: Evidence From An Online Labour Market" (2026). ECIS 2026 Proceedings. 4.
https://aisel.aisnet.org/ecis2026/platforms/platforms/4
Exploring The Effect Of Platform-Embedded Generative Ai On Skill Diversity: Evidence From An Online Labour Market
Generative AI (GenAI) is reshaping online labour markets. Existing research mainly examines general-purpose GenAI, such as ChatGPT, and focuses on aggregate outcomes, including falling demand and compressed prices in easily automated tasks, while revealing little about the demand for work skills and the role of platform-embedded GenAI. We explore how platform-embedded GenAI affects work skills. Leveraging logo design job posts before and after the launch of an early-stage platform-embedded logo-AI tool on the online labour market EPWK, we investigate its effect on requested skill diversity using a difference-in-differences design and a new large language model-based skill extraction and embedding framework. We find that logo jobs exhibit higher skill diversity than other design jobs after the platform introduced logo-AI, and that stronger competition among freelancers partially mediates this effect. Our findings indicate that changes in skill demand in online labour markets are an outcome of introducing platform-embedded GenAI.