Generative Artificial Intelligence: A PRISMA-Based Systematic Review of Models, Applications, Risks, and Future Research

Abstract

Generative Artificial Intelligence (GenAI) is one of the most impactful technological developments of the past five years. They have disrupted research, education, healthcare, software engineering, business, cybersecurity, and creative industry sectors. Unlike predictive AI, generative models learn data distributions to create new content like text, code, audio, images, multimodal artifacts, and synthetic data. The present research paper is a systematic review based on PRISMA, in which 50 candidate studies were identified between 2021 and 2026 and selected for final verification in the databases of IEEE Xplore, Scopus, Web of Science, and major publishers. The review examines the leading family models, areas of application, benefits and risks, and future research. It is found that the major technical foundations of present-day GenAI research are large language models, diffusion models, generative adversarial networks, and multimodal foundation models. GenAI promises to enhance productivity, personalization, software automation, knowledge access, medical communication, and synthetic data creation; however, it can also raise serious concerns about hallucination, bias, privacy, copyright, misinformation, cybersecurity misuse, academic integrity, explainability, and governance. The conclusion of the paper suggests that quality GenAI research has to go beyond merely demonstrating performance to transparent evaluation, domain-specific validation, human oversight, source attribution, and responsible governance.