Artificial Intelligence Applied To Consumption Forecasting In The Energy Mix: A Bibliometric Analysis

The global shift towards sustainable energy systems has positioned artificial intelligence as a decisive tool for optimizing generation, prediction, and uncertainty management in renewable energy sources. This study presents a bibliometric analysis of the integration of Artificial Intelligence methods in renewable energy microgeneration, with a specific focus on electricity consumption prediction and energy mix optimization. A total of 1,684 documents indexed in Scopus from 2003 to 2026 were examined using the Science Mapping Workflow methodology and tools, including Bibliometrix and RStudio, for network analysis and thematic evolution. The analysis reveals exponential growth in the scientific corpus with an average annual rate of 23.7% until 2024. Geopolitically, there is a shift in scientific leadership towards Asia, with China and India in the lead, followed by the United States. The conceptual structure identifies a technological transition from deterministic control models to deep learning architectures and recurrent neural networks, which better manage the non-linearity and stochasticity of renewable sources. In this sense, Artificial Intelligence has gone from being a complementary technology to a central element of digital energy infrastructure.