Deepfake: An Endanger to Cyber Security

‘Deepfake’ got originated from the technology ‘deep learning’ working behind it and the type of information it generates ‘fake’ after manipulating the original information. It’s an AI-based innovation used to make counterfeit recordings and sound that look and sound genuine. The inventions and implication of digital technologies in all spheres of mankind, has posed a great challenge to mankind to come up with secure solutions against a digital problem called Deepfake resulting from the application of deep learning thereby compromising authentication or originality. The said technology creates digital images and videos all new and totally fake. But the consequence it creates on society is totally a negative impact. With the aid of AI in developing hyper realistic videos, Deepfake has extended its giant wings to harm societal health and resulting in a critical challenge against the authenticity of the source. The Internet has become the platform to deliver these Deepfakes to unlimited destinations within no time. There are lot many researches have been carried on how to detect this deep fakes. Most of the research works have used deep learning models like Convolution Neural Network (CNN) for analyzing the convolution traces in deepfakes. Some of the research works have used Recurrent Neural Networks (RNN) by combining the Long Short-Term Memory (LSTM) with Blockchain. This research study has presented the comprehensive literature study, which highlights the various approaches used in generation and detection of deepfakes.