Deepfake Detection on Social Media: Leveraging Deep Learning and Fast text Embeddings for Identifying Machine-Generated

Abstract

Deepfake technology, which uses AI to create manipulated media, poses a significant threat to information integrity on social media platforms. In India, the rise of deepfake content has grown exponentially, especially in the political and entertainment domains, where fake news and AI-generated videos have gone viral, leading to misinformation.  The primary objective is to develop a robust AI 
model that accurately detects deepfake content on social media platforms, focusing on identifying machine-generated tweets using FastText embeddings. Traditional methods involved human moderation, fact-checking agencies, and manual filtering of social media posts based on predefined rules and keyword matching.