The usage of ATM’s has increased in huge numbers. As technology continues to develop rapidly, conventional ATMs are assailable to theft. It is no secret that computer vision is advancing rapidly in today's world. As biometric identification techniques have advanced over the past few years, including fingerprinting, retina scanning, and facial recognition, the past few years have seen an increase in the number of security measures at ATMs to increase security. Specifically, this project aims to give a computer vision method to solve the security risk associated with accessing ATMs. This proposes an ATM security model that uses electronic facial recognition using Deep Convolutional Neural Network. where faces would be protected as well as their accounts. Face Verification Clickbait Link will be generated and sent to the bank account holder to verify the identity of the unauthorized user through some dedicated artificial intelligent agents, for remote certification. As the biometric features cannot be cloned, this proposal will make it possible for the account holder to be the only individual who has access to their account. By using a real-time data set, fraud stemming from ATM card theft is eliminated.