Modern digital cameras are susceptible to producing noisy images in low light conditions due to sensor limitations. This noise manifests as randomly colored pixels that appear as grain, degrading overall image quality. Traditional denoising algorithms have limitations in differentiating actual image detail from noise. Recent advancements in artificial intelligence (AI) and deep learning offer potential new approaches for more intelligently identifying and reducing image noise while preserving real details and edges. In this paper, we propose a novel AI-assisted denoising algorithm that leverages a convolutional neural network (CNN) to differentiate noisy pixels from those containing actual image signal. The trained model classifies each pixel as either signal or noise. Pixels identified as noise are smoothed by averaging the color values of neighboring signal pixels, while pixels classified as true signal are left unaltered to maintain sharpness. We evaluate our algorithm on a dataset of noisy raw images from various camera sensors under low light conditions. Both objective quality metrics like Peak Signal-to-Noise Ratio (PSNR) and subjective human ratings demonstrate that our AI-based denoiser outperforms leading traditional denoising methods, especially in terms of preserving real image details and edges while smoothing away noise. Our solution has applications in computational photography, image processing pipelines, and may help overcome hardware limitations of small camera sensors in portable devices like smartphones.