AI Model Cards
Detailed transparency documentation for our assistive AI systems. Designed for accountability, mapped to DHS grant requirements, and built with human oversight at the core.
Entity Matching Assistant
Intended Use
Assist reviewers by clustering likely matches across sanctions sources
Visitor name and context signals (e.g., aliases, locale)
Match candidates with similarity signals, not determinations
Human-in-the-loop adjudication required
Ambiguous names; transliteration variance
Risk Summary Assistant
Intended Use
Summarize relevant risk context for reviewer attention
Candidate matches and associated metadata
Short risk summaries highlighting key factors
Human approval required; summaries are advisory only
May omit non-critical details; relies on source quality
ID Vision Parser (Optional)
Intended Use
Extract text from ID images when enabled by org policy
ID photo captured with consent
Parsed text fields for faster form fill
Human validation; disabled by default
Image quality, glare, occlusion
These model cards are living documents, updated as our systems evolve. Last updated: March 2026