A fintech AI project uses third-party data sources for credit risk modeling.

PMI’s trustworthy AI framing highlights governance, transparency, and accountability as essential ingredients for systems people can interpret and monitor. When third-party data feeds can change, the PMI-aligned approach is to establish governance and supplier controls that define data quality expectations, lineage, permitted uses, privacy constraints, and monitoring/audit mechanisms. This supports accountability by making data dependencies explicit and enabling early detection when upstream changes degrade model behavior. Removing external data (B) may be unnecessary and can reduce predictive power; a responsible approach is controlled use, not blanket elimination. One-time documentation at launch (C) fails to address lifecycle change. Allowing inconsistent definitions across teams (D) increases risk of aggregation errors and noncompliance. PMI-CPMAI’s emphasis on responsible practices (privacy/security, governance, monitoring) supports the structured governance and monitoring option as the best control.