Skip to content
V Vulcan Industrial AI Program · Proposal

AI Basics · Plain English

Data Classification

Sorting information by sensitivity — Public, Internal, Confidential, Restricted — so every AI decision starts from the data, not the tool.

Data classification is the discipline of labeling information by sensitivity before deciding what tools may touch it. This plan uses four classes: Public (anyone may see it), Internal (everyday business content), Confidential (commercially sensitive), and Restricted (export-controlled, defense-related, or customer-controlled).

Every platform decision in this proposal flows from those labels: Internal content may use tenant-controlled cloud AI; Restricted content only ever touches local models. Without classification, AI governance is guesswork; with it, the rules are mechanical enough to audit.

Where it shows up in this proposal: the four classes anchor the Security and Governance sheet, and every department card on AI by Department carries a data-sensitivity note.