Legal footing
This extends mechanisms that already exist in federal law rather than inventing a new authority: the Child and Family Services Reviews, the Citizen Review Panels that CAPTA already requires every state to convene, and the federal outcomes survey (NYTD). It gives those existing bodies the cross-referenced, life-long evidence they were always meant to have and never had.
Consent, for grown survivors
An adult who aged out opts in. Their consent to link their adult self back to their childhood token is what makes this both ethical and legal. They control their own story and can withdraw it. Testimony is gathered in a trauma-informed way, participants are compensated for their time, and they are treated as the experts they are rather than as data points.
The present-danger door
If a grown survivor reveals that an abuser is fostering children today, pure anonymity would stop anyone acting — and that is unacceptable. So there is one narrow, firewalled path: it requires the key-holder quorum to convene, a documented and legally authorised reason, and a permanent record of who asked, why, and what they saw. In normal operation this door stays shut. It exists to protect a living child, and for nothing else.
Protection from retaliation
Coming forward can invite retaliation. The tokenisation is itself the strongest available shield: a participant's identity is never in the auditable record to begin with. Findings are published as patterns, never as anything traceable to the person who lived them.
Who governs it
A governance board with former foster youth and survivor-parents on it — not as a courtesy seat, but with real authority over how the system is used and what is published. The people who lived it decide how their evidence protects the next child.
The rule the pattern engine runs under
The engine surfaces patterns for human review; it never renders a verdict. It does not declare that anyone lied. It flags where the record contradicts itself, where evidence went missing, or where one person’s cases are a statistical outlier — and hands that lead to a human being to investigate. “The record shows a contradiction here” is durable in a way “the algorithm says she lied” never could be, and it protects real people from being branded by software.
What the pattern engine looks for
| Pattern | What it means |
|---|---|
| Whole-life outcome patterns | Which childhood circumstances actually led to which adult outcomes — measured from real lives, not from what a closing summary claimed. |
| Wrongful-removal patterns | Where a child was taken on an allegation the record itself later contradicts. |
| Missing-or-altered-evidence patterns | Evidence logged and then gone, or records changed around key dates. |
| Actor-cluster patterns | A repeat player whose cases cluster with unusually bad outcomes or later-proven irregularities. |
| Cross-generational patterns | How harm echoes forward — visible when an aged-out adult’s real story is matched back to the child they were. |
Where this is, honestly
| Phase | What happens | Status |
|---|---|---|
| Phase 1 — Foundation | Stand up the subdomain and Ally; publish the aggregate measures already available from AFCARS, NCANDS and NYTD. | in progress |
| Phase 2 — The Vault | Build the tokenization engine and the vault to the Section 5 architecture. Independent security audit before any real data enters. | not started |
| Phase 3 — Cross-referencing | Bring datasets in one at a time under the token, each behind a data-sharing agreement. | not started |
| Phase 4 — Testimony | Open the consented survivor pathway; link adult outcomes back to childhood tokens. | not started |
| Phase 5 — Patterns | Integrate pattern recognition on the assembled data, with human-review-not-verdict enforced throughout. | not started |