How can we automate attribution across locales when an AI answer cites multiple hub facts, and how should this attribution be presented to users?

Create a provenance panel in the Digital Twin UI that lists each locale fact_id used, the proofs consulted, and a concise attribution summary (which location contributed which sources). Tie this to a Change Log entry showing localization updates behind the citations. Keep internal graph structures private and display only canonical facts, proofs, and language mappings. This enables credible, auditable attribution across locales while preserving privacy and protecting the underlying data architecture.