Ensemble Deep Learning Approaches for AI-Altered Video Detection

arXiv:2607.06872v1 Announce Type: new Abstract: The increasing accessibility of artificial intelligence has led to a rapid rise in AI-generated videos, making it more difficult to distinguish between real and manipulated content. Many existing detection methods rely on a single model and often struggle to generalize across different types of deepfakes. In this work, we developed a multimodal deepfake detection system that combines both audio and visual analysis using an ensemble of models. The s...