AIGen Automates AI Bill of Materials Generation for MLOps Compliance.

The responsible development of AI systems necessitates thorough documentation of all their components, from datasets to model weights and dependencies. While standards like SPDX 3.0 now support AI and dataset profiles, practical tools for automated, standards-compliant AI Bill of Materials (AIBoM) generation have been lacking. AIGen addresses this gap by providing a modular AIBoM generator. It produces machine-readable inventories of AI system components that adhere to the SPDX 3.0 AI profile. The tool integrates with the MLflow MLOps framework and employs a combination of mining heuristics and Large Language Models to create AIBoMs. Designed with extensibility in mind, AIGen features a plugin interface, allowing developers to add domain-specific collectors for various AI frameworks like Hugging Face, PyTorch, and TensorFlow without altering the core codebase. This innovation is crucial for facilitating compliance with major AI regulations such as the European Union AI Act, the NIST AI Risk Management Framework, and ISO/IEC 42001, offering a foundational solution for transparent AI supply chain governance.