Theses and Dissertations

Date of Award

7-1-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Information Systems

First Advisor

Bin Wang

Second Advisor

Geng Sun

Third Advisor

Francis Kofi Andoh-Baidoo

Abstract

Digital platforms increasingly incorporate automated bots and generative artificial intelligence (GenAI), reshaping the signals participants use to judge which content spreads, which freelancers are hired, and which contributors emerge as opinion leaders. Yet when automation and GenAI can reproduce these signals, it becomes unclear whether they continue to credibly signal quality and value. This dissertation addresses that gap across three platform settings , each treated as a signaling environment. The first study compares the virality of human- and bot-generated Bitcoin tweets. Using zero-inflated negative binomial regression on 15,762 tweets classified by bot score, it finds that bot-authored tweets are less likely to be retweeted, though linguistic features differ by account type: syntactic complexity helps bots more, while sentiment polarity and lexical richness benefit human posts. The second study turns to an online freelance labor market. Using conditional logistic regression with project fixed effects on 47,701 Freelancer.com bids, it finds that explicit GenAI references in bid text lower the probability of being hired, while price, delivery time, and reputation remain dominant. The third study examines opinion leadership on Stack Overflow. Using a twenty-quarter event-history panel and treating the November 2022 ChatGPT release and AI-answer ban as a joint shock, discrete-time logit and Cox models show that signals predicting leadership entry shift structurally, with peer-validated quality strengthening as accumulated recognition weakens and a smaller, faster pathway emerging. Taken together, automation and GenAI reconfigure rather than erase platform signals, shifting value toward cues that are harder to fake, distinctly human, and peer validated. Theoretically, the dissertation extends signaling theory to GenAI-shaped settings by showing that signal credibility depends jointly on production technology and platform governance, linking social bots, online labor markets, and knowledge communities within a common signaling framework. Practically, the findings suggest that platforms should preserve the credibility of the signals that allocate attention through content diffusion, opportunity through hiring, and status through opinion leadership, while participants build durable influence through human and peer-recognized contributions that are difficult for machines to imitate.

Comments

Copyright 2026 Chaitali Bonke. All Rights Reserved. https://proquest.com/docview/3409002825

Download

Share

COinS