Definition
AI enrichment in synthetic data is the process of using a language model to add contextual, narrative, and culturally specific details to profiles whose numerical and structural fields have already been established through mathematical generation. This includes generating realistic names, company affiliations, industry descriptions, wealth source narratives, and other text-based attributes that require cultural and linguistic knowledge. AI enrichment operates as a second-stage process, layering human-readable context onto a mathematically sound foundation without altering the underlying numerical values.
Why It Matters for Synthetic Data
Pure mathematical generation produces numerically accurate profiles but leaves them impersonal — Profile #4,721 with $87M net worth and a 60/25/15 asset split tells a model nothing about cultural context, naming patterns, or industry affiliations. For compliance testing, onboarding simulation, and AI training, these contextual details matter. KYC systems must handle Arabic patronymics, East Asian naming conventions, and European compound surnames. Risk models need industry context to assess wealth source plausibility. AI enrichment bridges the gap between mathematical accuracy and real-world usability.
How Sovereign Forger Handles This
AI Enrichment is Stage 2 of Sovereign Forger’s pipeline. After the Math-First stage locks all numerical fields, a locally-hosted Qwen 32B language model generates culturally appropriate names (using onomastic rules tied to each of the 31 archetypes), company names, industry affiliations, and wealth source descriptions. Critically, the LLM runs entirely offline on local hardware — no record ever touches a network or external API. The enrichment is constrained by archetype parameters: a Middle Eastern sovereign family archetype generates Gulf-appropriate names and petroleum/investment industry affiliations, while a Pacific Rim semiconductor dynasty archetype generates East Asian naming patterns and tech sector context. The AI adds narrative depth; it never overrides mathematical truth.
Related Terms
FAQ:
Q: What is AI enrichment in simple terms?
A: It is using an AI language model to add realistic names, company details, and cultural context to synthetic profiles that already have their financial numbers established.
Q: Does AI enrichment compromise data privacy?
A: Not when implemented correctly. In an offline-first architecture, the AI runs locally and generates contextual details from learned patterns rather than referencing real individuals. No PII is input, queried, or transmitted during the enrichment process.
