Dissertations, Theses, and Capstone Projects
Date of Degree
9-2024
Document Type
Capstone Project
Degree Name
M.A.
Program
Digital Humanities
Advisor
Michelle A. McSweeney
Subject Categories
Digital Humanities
Abstract
Large Language Models (LLMs) have made remarkable advancements, reaching a level where they can generate outputs that compete with human writing. LLM's sophisticated features, however, make it hard for us to spot the difference between human-written and AI-generated text. LLMs have several weaknesses since they are susceptible to hallucinations and outputting factually incorrect information. Additionally, they may be used maliciously to promote deception and instigate fraud, such as fake reviews generation, which rapidly appear on the internet while posing a challenge for consumers who want to place their faith in online reviews. To maintain trust in the online information system, it is essential to discover methods to detect false reviews. This capstone project aims to construct a classifier to identify fake reviews accurately.
Recommended Citation
Dey, Zico Abhi, "Assessing the Influence of LLMs on Online Review: Identifying and addressing the Rise of Convincing Fake Reviews" (2024). CUNY Academic Works.
https://academicworks.cuny.edu/gc_etds/6199
Colab Notebook - LLM Text Generation
LLM_text_generation_notebook.py (2 kB)
Code - LLM Text Generation
