A foundation model is a very large machine learning model trained on broad, diverse datasets to learn general patterns that can be adapted to many downstream tasks. After initial training, the model is adapted through fine-tuning, prompt engineering, or retrieval methods to solve specific problems such as summarization, classification, or code generation.
Foundation models are valued for their flexibility but require careful governance, safety checks, and technical controls to ensure reliable, fair, and cost effective use in production environments.
