Detecting AI Generated Content in Short Form Text
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
As the use of AI-powered language models continues to grow exponential-ly, the ability to distinguish human generated text from AI generated text has become critical to combat misinformation and uphold trust on online platforms. This study addresses the novel challenge of classifying casual, short-form user reviews on e-commerce websites as either human generated or AI generated. A custom dataset of over 60,000 authentic reviews and 17,645 AI generated reviews across diverse product categories was engi-neered by web scraping and prompting state-of-the-art large language mod-els. Through rigorous feature engineering, model development, and statisti-cal analysis, this research demonstrates the feasibility of accurately dis-criminating between human generated and AI generated content in informal contexts.
