This paper presents a machine learning-based approach for detecting artificial and deceptive faces. The method combines advanced facial recognition techniques with custom convolutional neural networks (CNNs) to enhance the accuracy of face authentication. The system uses deep learning to analyze facial features and detect anomalies that indicate fake or manipulated faces. The approach leverages the uncanny valley phenomenon to identify fake faces, using multi-task learning to determine the authenticity of a face and predict its future state without explicitly simulating human brain processes. The proposed method improves upon traditional feature-based approaches by utilizing deep learning to analyze hierarchical representations of input data. It employs CNNs and generative adversarial networks (GANs) to extract complex visual features and detect fake faces. The system is designed to be robust against various forms of face manipulation, including deepfakes and other advanced techniques. The method has been tested on a dataset containing 30,000 real and 30,000 fake face images, achieving a median precision of over 99.4%. The system includes several modules for face recognition, detection, preprocessing, and feature extraction. It uses Gabor wavelets for feature extraction due to their biological relevance and computational capabilities. The CNN-based approach is trained to distinguish between real and fake faces, with each face group having its own model. The system also includes an edge detector to identify borders in fake faces, using a convolutional matrix to detect maximum transitions in different directions. The proposed method has been validated through extensive experiments, demonstrating its effectiveness in detecting fake faces. The results show that the system can accurately distinguish between real and fake faces, even in the presence of advanced manipulation techniques. The method is designed to be adaptable and can be integrated into various applications, including fraud detection and face recognition systems. The system's performance is evaluated using metrics such as accuracy, precision, recall, and F1-score, ensuring its reliability and effectiveness in real-world scenarios.This paper presents a machine learning-based approach for detecting artificial and deceptive faces. The method combines advanced facial recognition techniques with custom convolutional neural networks (CNNs) to enhance the accuracy of face authentication. The system uses deep learning to analyze facial features and detect anomalies that indicate fake or manipulated faces. The approach leverages the uncanny valley phenomenon to identify fake faces, using multi-task learning to determine the authenticity of a face and predict its future state without explicitly simulating human brain processes. The proposed method improves upon traditional feature-based approaches by utilizing deep learning to analyze hierarchical representations of input data. It employs CNNs and generative adversarial networks (GANs) to extract complex visual features and detect fake faces. The system is designed to be robust against various forms of face manipulation, including deepfakes and other advanced techniques. The method has been tested on a dataset containing 30,000 real and 30,000 fake face images, achieving a median precision of over 99.4%. The system includes several modules for face recognition, detection, preprocessing, and feature extraction. It uses Gabor wavelets for feature extraction due to their biological relevance and computational capabilities. The CNN-based approach is trained to distinguish between real and fake faces, with each face group having its own model. The system also includes an edge detector to identify borders in fake faces, using a convolutional matrix to detect maximum transitions in different directions. The proposed method has been validated through extensive experiments, demonstrating its effectiveness in detecting fake faces. The results show that the system can accurately distinguish between real and fake faces, even in the presence of advanced manipulation techniques. The method is designed to be adaptable and can be integrated into various applications, including fraud detection and face recognition systems. The system's performance is evaluated using metrics such as accuracy, precision, recall, and F1-score, ensuring its reliability and effectiveness in real-world scenarios.