Child privacy
Child voice data deserves clear rules.
A child's voice is sensitive data. This page explains—in plain language—who manages the child account, what may be recorded, who can access it and how it is deleted. For parents, clinics and schools.
Clinician-managed accounts
A child account is created and managed by the treating clinician (or the school/clinic deployment)—not directly by a child.
Recorded only for practice
Short speech recordings are captured only when a practice exercise requires them, to produce structured feedback for the clinician.
Training is a separate choice
By design, therapy recordings are not used for model training by default. Any training use would require separate, explicit caregiver/clinic consent.
Designed to delete raw audio
The designed lifecycle deletes raw audio after analysis and review. Retention is an explicit clinician choice, never a silent default.
Deployment roles matter
Clinic and school deployments assign access roles: clinician, supervisor, admin, caregiver. Each sees only what their role requires.
Questions and deletion
Caregivers can request deletion or export through their clinician or by contacting [email protected].
Who manages the child account
Child accounts are created through a clinician or organization. Children do not create accounts directly, and we do not design the child experience to collect personal information from the child.
What may be recorded, and why
Only short audio of practice attempts for assigned target words, recorded to generate structured feedback (target sound, word position, observed pattern, confidence band). Recording quality checks run before analysis; poor recordings are flagged for a re-try instead of being judged.
Who can access it
- The treating clinician (recordings, structured findings, progress).
- The linked caregiver, at the visibility level the clinician/clinic configures (completion, general feedback).
- Supervisor/admin roles only where the deployment enables them, and only at the level their role requires.
Retention
Raw audio is designed to be deleted by default after analysis and any clinician review. Structured results follow the clinic’s retention policy. Exact default periods for child recordings will be published on the Security page before the clinical pilot begins.
Training choice
By design, child recordings are not used for model training by default. Where a separate opt-in program exists in the future, it will require explicit caregiver/clinic consent, use de-identified data, and allow withdrawal at any time.
Parent and guardian controls
Caregivers can see what was assigned, what was completed and what was sent for review—without needing to score their child's speech. They can request deletion or export of their child's data.
School and clinic deployments
When Articu is deployed by a school, the institution may act as the controller of education records under applicable law (for example, FERPA in the United States); the contractual roles and data handling are set out in the deployment agreement. We do not claim certifications that have not been verified.
Contact
Child data questions: [email protected].