Neural Networks as Artists: Exploring AI in Contemporary Painting

Abstract

The integration of neural networks into contemporary painting represents a profound transformation in artistic production, aesthetic theory, and creative authorship. This study examines how artificial intelligence (AI), particularly deep learning architectures such as convolutional neural networks (CNNs), generative adversarial networks (GANs), and diffusion models, functions within modern painting practices. Employing an interdisciplinary IMRAD framework, the research combines technical analysis of neural architectures, case studies of AI-generated artworks, and theoretical evaluation of creativity, authorship, and aesthetic agency. The findings indicate that neural networks enable stylistic emulation, generative image synthesis, semantic text-to-image translation, and large-scale visual recombination, thereby expanding both the visual and conceptual boundaries of painting. However, their outputs reflect procedural and statistical creativity rather than conscious intentionality. The study further demonstrates that AI reshapes artistic workflows by shifting emphasis from manual execution toward prompt engineering, system configuration, dataset curation, and algorithmic mediation. While AI-generated paintings have achieved institutional validation and market recognition, they simultaneously raise significant ethical and legal concerns regarding copyright ownership, dataset consent, artistic labor displacement, and cultural appropriation. Ultimately, neural networks function not as autonomous artists but as transformative creative media that redefine collaboration between human imagination and machine computation in contemporary art.