Automated AI Model Development: a Systematic Literature Review

Artificial intelligence (AI) has rapidly evolved, presenting both opportunities and challenges in the development and deployment of sophisticated models. Traditional methods for creating AI models often require extensive human expertise for tasks such as feature engineering, hyperparameter tuning, and model selection, making the process time-consuming and prone to bias. In response, automation techniques-including automated machine learning (AutoML), hyperparameter optimization, and neural architecture search-have emerged to streamline model generation. This systematic literature review explores current trends, benefits, and challenges associated with AI model automation. Drawing upon articles published from 2020 onwards, the review follows the CIMO (Context, Intervention, Mechanism, Outcome) framework and adheres to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines to ensure methodological rigor. Searches were conducted in Scopus, Web of Science, and Academic Search Ultimate, focusing on peer-reviewed studies that address AI model automation in supervised, unsupervised, and reinforcement learning domains. Each study's context, intervention techniques, underlying mechanisms, and outcomes were extracted and assessed for quality and relevance. By synthesizing the findings of recent research, this review not only highlights advancements in automated AI model development but also identifies gaps in existing knowledge. The results provide critical insights for researchers and practitioners, guiding future exploration of scalable, efficient, and reliable AI automation strategies.