LLM match conditions in Sensitive Data Scanner
Sensitive Data Scanner can use a natural language prompt and examples for LLMs to evaluate datasets, virtual tables, and media sets for sensitive data.
Read Palantir's release noteshttps://www.palantir.com/docs/foundry/announcements/2026-09/#detect-sensitive-data-using-language-model-match-conditions-in-sensitive-data-scanner
Summaries of vendors' own notes. Product names and logos belong to their owners; logos via logo.dev.
More Foundry and AIP releases
Also shipped on Sep 10, 2026
BigQuery ML.METRICS evaluates predictions without a stored model
PreviewSQLGovernanceObservability
BigQuery Data Engineering Agent integrates with BigQuery Graph
GAAI agentsSQLData integration