Never enter this here
Identity and contact details
Do not enter a child's full name, phone number, email, account identifier or identity document.
Sensitive-user safety boundaries
Protect the person before continuing the prompt.
This public guide says what must not enter an AI flow, when the flow must stop, and when a responsible person must take over. This guide has no input form and sends no personal report.
AI identity disclosure
Declared AI operating roles, not human employees.
Declared accountable AI unit: Legal and Risk. Declared independent AI checker: Quality Control.
The actual runtime creator and model version are not verified. These declared roles have no spending or external-action authority in this public record, and this disclosure does not claim whole-product coverage.
Inspect the AI identity record5current official references
8input categories kept out
6conditions that stop the AI flow
0write, consent or crisis-routing endpoints in this guide
Before typing
Keep identifying and high-risk information out of the prompt.
These are data-minimization boundaries for every user. They are especially important when a child or a vulnerable person may be involved.
Never enter this here
Precise location and routine
Do not enter a home address, live location, school, route, schedule or regular meeting place.
Never enter this here
Health and care details
Do not enter diagnoses, medication, treatment records, disability details or private care information.
Never enter this here
Intimate or sexual material
Never upload, request, create or share intimate images or sexual content involving a minor.
Never enter this here
Credentials and payment data
Do not enter passwords, recovery codes, payment details, bank information or account access tokens.
Never enter this here
Legal and case details
Do not enter identifiable case files, allegations, custody details or protected legal communications.
Never enter this here
Other people's information and secrets
Do not share another person's private information, confidential messages or a secret that creates pressure or isolation.
Never enter this here
Detailed crisis information
Do not use this page to submit a crisis report. In immediate danger, leave the AI flow and seek local human help.
Stop conditions
Some situations must leave the AI flow.
A fluent answer is not proof that the system understands risk, consent, urgency or the person's real circumstances.
- Stop 1There may be immediate danger, self-harm, violence or a medical emergency. Stop the AI flow. Contact local emergency services or a trusted adult who can act now.
- Stop 2The situation may involve abuse, exploitation, grooming, coercion or pressure to keep a harmful secret. Do not continue a private AI conversation. Move to a trusted adult or an appropriate qualified local service.
- Stop 3A request or file contains sexual content involving a person who is or may be a minor. Do not upload, create, edit or share it. Stop and seek appropriate local human help.
- Stop 4The answer would decide medical care, legal rights, money, safety or another high-impact outcome. Pause. Use a qualified professional who can verify the facts and take responsibility for the decision.
- Stop 5Age, consent, guardianship or authority to share information is unclear. Collect nothing and do not proceed until an authorized adult and the applicable process are established.
- Stop 6The AI answer cannot be verified, conflicts with evidence or is being treated as the only authority. Do not act on it. Verify with a trusted adult, original source or qualified professional.
Human handoff
The next step is a person, not another prompt.
Qeravio shows the boundary only. It does not identify, contact or transfer information to any person or service.
Guidance only
Trusted adult
Use when a child needs context, permission, protection or help checking an answer. Leave the private AI flow and speak directly with a parent, guardian, teacher or another responsible adult you trust.
Guidance only
Qualified professional
Use for medical, mental-health, legal, financial, educational-support or safeguarding decisions. Use a professional who can examine the real context, verify facts and accept responsibility for the advice.
Guidance only
Local emergency services
Use when someone may be in immediate danger or needs urgent medical or protective intervention. Stop using Qeravio for the situation and contact the verified emergency service for the person's current location.
Official reference map
Each source informs a boundary. None approves this product.
Every jurisdiction remains reference-only until verified product, operator and market facts receive qualified local review.
01
Global policy guidance
Guidance on AI and Children, version 3.0
A child-rights design reference. It does not approve Qeravio or decide local legal duties.
02
Voluntary United States risk framework
Artificial Intelligence Risk Management Framework
A general risk-management reference, not a child-specific certification or legal safe harbour.
03
United Kingdom
Age appropriate design code
A United Kingdom statutory-code reference. Qeravio's scope and duties have not been determined.
04
United States
Children's privacy and COPPA resources
A United States children's privacy reference. It is not a finding that Qeravio is covered or compliant.
05
Israel
Public recommendations for personal use of generative AI
A privacy-minimization reference for the public. It is not child-specific legal approval.
What is still missing
Published boundaries are not whole-product protection.
These gaps stay public and prevent any claim of child safety, legal compliance or live intervention.
- Gate 1Determine every target jurisdiction, age threshold, consent rule and child-directed-service test from verified operator and market facts.
- Gate 2Obtain qualified local legal review before making any compliance, suitability or age-assurance claim.
- Gate 3Complete native-human child-safety and safeguarding review with people qualified for the intended audience and use case.
- Gate 4Design, approve and independently test age assurance or consent only if a future product flow truly requires it.
- Gate 5Implement and verify whole-product enforcement, reporting and incident handling before claiming that these published boundaries are enforced everywhere.
- Gate 6Establish localized, staffed and verified crisis operations before adding any emergency routing or intervention claim.
Version binding
This safety review belongs to one exact Worker.
A code, content, policy or runtime change requires a new independent review. Evidence is never inherited.
- Evidence identifier
qeravio-qtop191-public-sensitive-user-safety-release-20260820T185710Z- Tested Worker
56bbaf52-e861-4c6d-870e-87c2369a2b78- Completed
- Aug 20, 2026, 6:57 PM
- Open review findings
- 0
- Public write endpoints
- 0
- Live interventions
- 0