Data Innovation Summit 2026: GenAI Synthetic Tabular Data – Variations vs Ontologies

Ericka Johnson presented at the Data Innovation Summit 2026 in Stockholm (May 6–8), arguing that synthetic data is not simply “more data” but actively reshapes how we understand and model reality.

Her session examined how ontological approaches can provide stronger structure and auditability for AI-driven synthetic data creation — moving beyond statistical accuracy metrics toward traceable, honest, and auditable data science practices. She outlined four key components for responsible synthetic data: provenance transparency, distribution auditing, downstream labelling, and testing for representation in edge cases and underrepresented groups.