Identity theft, facilitated by the advanced falsification techniques, spans various domains, posing significant threats to security and privacy. Despite biometric authentication systems to mitigate these risks, the ability of criminals to create convincing fake identities remains a pressing concern. Detecting AI-generated images is crucial not only for preventing fraud and safeguarding sensitive information but also for maintaining trust in digital interactions. In this paper, we propose a Facial Image Authenticity Verification (FIAV) method that integrates diverse No-Reference Image Quality Assessment methods merged with features describing various aspects of image’s visual quality, to effectively distinguish between authentic and AI-generated facial images. This paper underscores the importance of integrating advanced image quality assessment tools and leveraging AI capabilities to fortify defenses against sophisticated falsification techniques. By prioritizing data quality and adopting innovative approaches, we aim to enhance the integrity of biometric identity verification in the digital age. By addressing the challenges posed by modern identity falsification techniques, the proposed FIAV method demonstrates robust performance in discriminating between real and AI-generated facial images. To ensure the practical application and accessibility of our method, it was deployed on web interface, which simplifies user interaction and rigorously tests the model’s performance in real-world scenarios, providing instant feedback on the classification of new images (available at https://github.com/yessin5/Classification-of-Real-and-Fake-face-images-based-on-image-quality-features.git ).