Traditional paper money and contemporary electronic money are two significant forms of exchange. However, due to the new and improved methods that counterfeiters are using, it is now becoming an increasingly important issue. This work suggests a machine-assisted method that has been created to distinguish between fake and real currency. Generative adversarial networks (GANs), a cutting-edge machine learning model, are used for this purpose. To develop a model that can be used for supervised prediction, GANs utilize semi-supervised learning. Semi-supervised learning is an exceptional approach in building a model that is fit for supervised prediction. It offers a multitude of benefits, allowing the training of unlabeled data while achieving highly accurate predictions. In this work, detecting counterfeit money using GANs involves a multi-step methodology that leverages the power of deep learning for accurate identification. Preprocess the data by resizing, normalizing, and augmenting the images to ensure uniformity and improve model generalization. The discriminator minimizes the loss when making correct classifications, while the generator minimizes the loss when the discriminator classifies its generated images as genuine. Indian banknotes were subjected to this procedure. Modern feature identification and image processing techniques were employed to develop the overall strategy of a valid input. The trials demonstrate that a high-precision system may be created to identify real paper money. During the training process, the generator network and the discriminator networks update their loss functions. Once training is complete, the model can accurately identify real money notes from fake ones with an 80% accuracy rate.