Anomaly Detection of News Articles through Agentic Artificial Intelligence; Influence of Social Media on American Presidential Elections 151 views

In today’s digital world, the whole world is connected through media and perhaps even becoming dependent on it for communication and for learning about what is happening in the world. My capstone project was about building a software system that used Agentic Artificial Intelligence (AI) to discover anomalies in news sources. The reason for undertaking this research was to help government agencies discover dangerous anomalies like a disease outbreak, economic crisis, and political unrest as soon as possible. My Science, Technology, and Society (STS) paper concerned the impact of social media on American presidential elections. The reason for undertaking this research was due to the increased importance of the role of social media in the latest presidential elections. Both the technical capstone project and STS research paper look into online media sources. These online media sources are having a growing impact and are becoming places where people depend on to find information whether that is discovering anomalies or learning about upcoming elections. My capstone project was creating a software system that uses Agentic AI to discover anomalies from news sources. The problem that this project offers a solution to is the arduous task of looking through several news sources and articles to find impactful anomalies that could warrant government agencies to take action in threatening situations, whether that be an outbreak in disease, early signs of economic downturn, or political unrest. My capstone project automates this arduous process and uses the cutting edge technology of Agentic AI. The method in which the capstone project achieves this is through an Agentic AI network that consists of several AI agents with specific tasks to break down this process. They receive data to analyze from a data pipeline that gathers news articles from online sources. When the Agentic AI network discovers an anomaly, the anomaly is sent to the project’s website for users to view. At the completion of the project, the software produced several impactful anomalies at a consistent, yet not overwhelming rate. The primary goal of finding anomalies from online sources was achieved to a satisfactory level. The project has the potential to free up employees that might be stuck having to do this tedious task, making their jobs a lot easier. It was determined however that more testing should be done to ensure the quality of the anomalies being produced. My STS research was about looking into the impacts of social media on American elections. The research question is to identify how social media platforms influence the outcomes of presidential elections in the United States beginning from the 2008 election. This topic is significant as the usage of social media has increased among not only presidential candidates but also voters. The methodology for this project was to study each individual election from 2008-2024 and find any trends among the five case studies. Online papers were the primary source for evidence for analyzing the five elections. Results from this project showed that not only has social media usage increased but the number of social media platforms used for elections has increased as well. Additionally, whichever candidate is more effective at using social media seems to have an edge in the elections. The conclusion of this research was that social media does play a large influence on the outcome of presidential elections but that it is important to remember that it is simply just one influencing factor in many when it comes to American presidential elections.

BS (Bachelor of Science)

Agentic Artificial Intelligence; Social Media; Election

School of Engineering and Applied Science Bachelor of Science in Computer Science Technical Advisor: Rosanne Vrugtman STS Advisor: Pedro Augusto Francisco

2026-05-08