Responsible AI
Defined purpose. Human oversight. No autonomous clinical authority.
Effective date: 12 July 2026 | Version 1.0
Contents
1. Purpose and boundaries
This statement explains the boundaries VitaPing applies to AI-enabled features and TGMNI discussions. It describes current capability and required controls without implying certifications, audits or deployments that have not occurred.
Core boundary: Insight support, never autonomous clinical authority.
2. What AI may support
Where enabled and approved, AI may assist with structuring information, checking completeness, organising authorised notes, producing reviewable summaries and helping approved institutions explore de-identified patterns.
3. What AI must not do
VitaPing AI must not diagnose, prescribe, determine treatment, replace emergency triage, make autonomous safeguarding decisions, score a person’s worth or risk, decide insurance or employment eligibility, or remove meaningful human review.
4. Human review
A person remains accountable. AI-assisted output must be identifiable as assisted, reviewable by an authorised person and capable of being corrected or rejected. High-impact use requires a named institutional owner and escalation route.
5. Data boundaries
AI processing must follow data-minimisation, access-control and retention rules. Personal emergency information must not be reused to train general-purpose models or for unrelated research without a separate lawful, transparent and approved basis.
6. TGMNI governance
Each TGMNI use case must define the institutional partners, purpose, legal roles, source data, de-identification standard, permitted outputs, model or analysis tools, human review, bias considerations, security controls and exit conditions.
7. Testing and change
Before deployment, the responsible team must test intended functionality, known failure modes, security, usability, relevant subgroup performance and human-review effectiveness. Material changes require review and controlled release.
8. Transparency
When an AI-enabled feature is used, the relevant user or organisation should be told what the feature does, what information it uses, what it cannot do and how to raise a concern.
9. Contact
Raise an AI or governance concern by emailing privacy@vitaping.ae. Include the relevant organisation, date, feature and a description of the concern. Do not include unnecessary health information in an unencrypted email.
Related documents: