AI Data Bill of Materials (DBOM)
Govern and audit AI model training and inference datasets.
Use Case Summary
Just as a Software Bill of Materials (SBOM) lists dependencies, Bedrock’s Data Bill of Materials (DBOM) enumerates the datasets used for AI training and inference. Using lineage and classification data from the Metadata Lake, Bedrock automatically produces a verifiable record of every data source an AI model consumes — ensuring integrity, compliance, and explainability.
Key Capabilities
Automatic DBOM Generation
Track all datasets used for training, fine-tuning, and inference.
Lineage Proof
Provide traceable provenance for every data element in a model’s pipeline.
Audit Readiness
Simplify AI compliance and model risk assessments.
Value to Stakeholders
CAIO
Gain transparency into model training datasets for responsible AI.
CISO
Help ensure models are not poisoned with malicious data and that their data access pathways are secure.
GRC
Produce audit-ready documentation for AI governance frameworks.
CTO
Ensure technical alignment between data pipelines and compliance mandates.