Doctoral Dissertation Research: Navigating the Risks of Social Media Bots - A Study on Autonomous Agents and Human Cognitive Processing

Bots are increasingly prominent on social media platforms, engaging with users, generating content, and influencing public discourse. These AI-driven agents disseminate information with varied intentions, including shaping opinions and behaviors. Despite concerns about false information and polarization, the mechanisms through which bots influence human decision-making remain poorly understood.

This project investigates the influence of bots on public opinion about controversial topics such as genetically modified organisms (GMOs). By analyzing bot-generated content and its interactions with human users, the study examines how bots influence public discourse and human responses, with a focus on linguistic features such as polarization, politicization, and emotional framing.

Understanding these linguistic dynamics sheds light on how bots shape public opinion, amplify narratives, and influence the cognitive processes driving individuals’ reactions.

This research expands the computers-as-social-actors (CASA) framework to examine bots' influence on human decision-making, with a focus on both theoretical advancement and methodological innovation. Integrating observational and experimental approaches, it analyzes over 26 million tweets using advanced natural language processing (NLP) techniques, such as topic modeling and sentiment analysis.

The experiment employs LLM-simulated personas as human respondents, comparing their responses to those of real human participants. This approach not only tests bots’ influence on decision-making but also demonstrates the potential of LLMs in studying human cognitive processes and decision-making dynamics. Methodological innovations include a bot detection tool leveraging large language models (LLMs) and a sampling technique using LLM-generated personas.

These tools ensure robust insights into human-bot interactions while advancing the CASA framework’s predictive capacity. The findings deepen theoretical understanding of human-AI interactions, guide strategies to mitigate bot influence, and contribute to computational social science. Addressing critical societal challenges, this project aligns with NSF’s mission to advance science and promote societal well-being.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.