Peiyan Yu, Raveesh Mayya, Anindya Ghose (Major Revision at Information Systems Research)
We utilized a quasi-experiment to investigate the effects of enabling non-monetary virtual gifts on the streamers’ revenue outcome via paid virtual gifts and find a stimulating effect instead of cannibalization. Mechanism analyses suggest competition among users on gifting value and habit forming.
Best Student Paper Finalist, WISE 2022, Copenhagen
Peiyan Yu, Runshan Fu, Anindya Ghose (Preparing for submission)
In this paper, we study content creators’ social interaction effort decisions on a live-streaming platform with a Hidden Markov Model. We find that while social interactions under a live-streaming context are generally considered stimulating, these efforts bear costs and increase the likelihood of a streamer developing mental fatigue in the long-term. Moreover, streamers with a smaller loyal community base or lower-status that reflects overall devotion and capability suffer stronger exhaustion effects from social interactions. Our results suggest that platforms should consider developing tools to lower the streamers' cognitive cost of social interactions while maintaining audience engagement.
Peiyan Yu, Sonia Jaffe, Nicole Immorlica, Ted Liu, Kelly Monahan (Preparing for Submission)
This study investigates the impact of a generative AI tool on freelancers' job preferences and outcomes with a field experiment. Access to Microsoft Copilot, a GPT-based tool embedded in Microsoft Office, was randomly granted to half of the study sample on a US leading freelancer marketplace platform in October 2023. Our findings suggest freelancers realize different dimensions of productivity under real-life constraints and contractual designs and potential incentive misalignment for jobs that contract on time. For example, we find Copilot access leads treated freelancers to bid relatively lower hourly rate compared to other bidders. While high-earners shift to higher-value job postings, low-earners do not exhibit the same behavior. We develop a conceptual analytic framework that highlights the effort reduction dimension of productivity changes to rationalize the empirical findings.
Thi Mai Anh Nguyen, Peiyan Yu, Sonia Jaffe, Nicole Immorlica, Ted Liu (Work in progress)
While initial experimental observations indicate shifts in freelancers’ preference changes following the use of GenAI productivity tools, understanding the mechanism of changes and predicting the long-term distributional benefits of GenAI on bidding and hiring outcomes is crucial. This project extends the partial equilibrium freelancer job preference changes to build a general equilibrium model and answer the question how the broader market dynamics evolves as AI usage matures overtime.
Peiyan Yu, Joao Sedoc (Work in progress)
As social media especially live streaming platforms attract its users that seek sociality and the sense of community, introduction of Vtubers (Emobied by AI-generated avartar, voice, and script) may disrupt but induce new types of sociality. In this project we utilize the structures of transformers to model user community development process and characterize emerging sociality.
ICIS International Conference on Information Systems, Bangkok 2024
WISE Workshop on Information Systems and Economics, Bangkok 2024
CIST Conference on Information Systems and Technologies, Seattle 2024
Microsoft Economics Research Seminar, Boston, 2024
Symposium on Statistical Challenges in Electronic Commerce Research (SCECR), Lisbon 2024
NYC Joint PhD Student Colloquium, New York 2024
Marketing and the Creator Economy Conference, Columbia Business School, New York 2023
FLIP Seminar, NYU Stern 2023
CIST Conference on Information Systems and Technologies, Phoenix 2023
INFORMS Annual Meeting, Phoenix 2023
POMS Production and Operations Management Society, Orlando 2023
WISE Workshop on Information Systems and Economics, Copenhagen 2022
CIST Conference on Information Systems and Technologies, Indianapolis 2022
AOM The Annual Meeting of the Academy of Management, Seattle 2022