The growing influence of Artificial Intelligence (AI) in digital media has catalyzed the development of autonomous content creation, transforming the production and consumption of visual art, design, and cultural narratives. Despite advances in Generative Adversarial Networks (GANs), existing research often overlooks comprehensive frameworks that assess both the aesthetic quality and moral implications of AI-generated content. This research aims to bridge that gap by proposing a holistic evaluation framework tailored for GAN-generated visual media. The novelty lies in combining aesthetic feature analysis with moral assessment using deep learning and language-based evaluation tools. It utilizes the ArtBench-10 dataset, which consists of 10 diverse artistic styles. Preprocessing steps include z-score normalization, Contrast-Limited Adaptive Histogram Equalization (CLAHE), and resizing images to 256 × 256 pixels to standardize input and enhance visual clarity. For aesthetic analysis, key features such as color harmony, texture richness, symmetry, and compositional balance were extracted using the Histogram of Oriented Gradients (HOG) method. The proposed method employs a conditional GAN (cGAN) to autonomously generate artwork, while a GPT-based language model, along with a rule-based ethics filter, evaluates associated textual metadata to detect potential moral issues, such as bias, cultural appropriation, and offensive content. Implemented using TensorFlow 2.0, the model supports modularity and realtime evaluation. Results indicate superior classification performance, achieving 96.3% precision and a 95.2% F1-score, confirming the effectiveness of the proposed framework in selecting aesthetically coherent and ethically sound AI-generated visual outputs.