Automatic (false) video content production and creation is made possible by deep fakes thanks to techniques like generative adversarial networks. Deepfake technology is highly contentious because it can have far-reaching consequences for society, such as the manipulation of elections. Many studies have been conducted to find ways to identify deepfakes and lessen their potential harm. One strategy involves using neural networks and deep learning. In this article, we look at exception and Mobile Net, two deepfake detection technologies, as potential methods for automated discovery of deep fake movies via categorization tasks. We employ Face Forensics++’ straining and assessment datasets, which include four samples created with four widely used deepfake methods. Results indicate excellent accuracy across all datasets, with accuracy ranging from 91 to 98% based on the deepfake technologies used. We also created a polling system that uses a combination of all four techniques to identify phony movies.