by Luis Eduardo Muñoz Guerrero
ABSTRACT
In this study, we developed and implemented an artificial intelligence agent based on OpenAI’s GPT-4.0 technology for the detection of fake news on social networks. The agent analyzes trending profiles and topics using web scraping techniques and public data analysis to identify patterns of misinformation. In addition, cultural and geographic components were incorporated into the analysis, adjusting the model to account for factors such as linguistic variations, levels of digital literacy, and sociocultural differences in various regions of Colombia, including the Andean Region, Atlantic Coast, Pacific Coast, and the Eastern Plains. The results indicate that our agent outperforms traditional methods in terms of accuracy and speed, reaching an accuracy of 93% and an average time of 1.2 seconds per news item. These findings underscore the great potential of this agent for future applications in various social platforms and regional contexts.
