General Motors proactive governance approach combines global processes with technical automation across six core systems. Starting with governance and AI policies that define ownership and accountability, they built a stewardship community by domain, then deployed technical solutions that automatically collect metadata from GitHub, enforce quality checks for AI legislation, and enable stewards to verify classifications through intuitive workflows.
Key takeaways:
Proactive governance starts with clear policies defining roles, ownership, and expectations before deploying technology
Domain-based stewardship communities act as custodians who define business rules and data management standards
Automated metadata collection from GitHub repos, YML files, and model cards feeds into centralized catalogs at scale
AI legislation requires minimum data quality standards — GM automated ~20 essential checks across all AI/ML datasets
Data classification utilities use ML to predict classifications, then stewards verify through catalog workflows to reach 98% coverage
