AverCare is committed to ensuring that Artificial Intelligence (AI) systems operate under appropriate human oversight.
This Human Oversight Information Notice explains how AverCare maintains human involvement, accountability, review processes, and decision-making authority when Artificial Intelligence technologies are used within the AverCare ecosystem.
This document should be read together with:
Privacy Policy
AI Privacy Information Notice
AI Safety Information Notice
AI Transparency & Disclosure Policy
Responsible AI Framework
AI Ethics Policy
Human-in-the-Loop (HITL) Policy
AverCare believes that Artificial Intelligence should support human decision- making rather than replace it.
2. OUR HUMAN OVERSIGHT COMMITMENT
AverCare is committed to ensuring that:
Humans remain accountable for decisions affecting users
AI systems do not operate autonomously in clinical contexts
Patient safety remains the highest priority
Human reviewers can intervene when necessary
AI recommendations can be modified or rejected
Human oversight mechanisms remain effective throughout the AI lifecycle
3. WHAT HUMAN OVERSIGHT MEANS
Human oversight refers to the active involvement of appropriately qualified individuals in reviewing, validating, approving, modifying, rejecting, or escalating AI-generated outputs.
Human oversight helps ensure that:
AI outputs are appropriate
Clinical recommendations are reviewed
Risks are identified
Errors are corrected
Users are protected
AI systems remain aligned with their intended purpose
4. HUMAN-IN-THE-LOOP (HITL) GOVERNANCE
AverCare utilizes a Human-in-the-Loop (HITL) governance model.
Under this model:
AI System → Human Review → Clinical Validation → Action
This means that AI-generated outputs may be reviewed by humans before actions are taken, particularly when outputs may influence healthcare decisions.
Human reviewers retain authority to:
Approve outputs
Modify outputs
Reject outputs
Override outputs
Escalate outputs for additional review
5. HUMAN REVIEW OF CLINICAL AI OUTPUTS
Certain AI-generated outputs may require review by qualified healthcare professionals.
Examples include:
Clinical recommendations
Risk assessments
Health alerts
Triage recommendations
Treatment pathway suggestions
Medication-related information
Chronic disease management recommendations
Mental health risk assessments
Fertility-related assessments
Pediatric health recommendations
The level of review depends on the nature and risk of the output.
6. HUMAN OVERRIDE CAPABILITIES
Authorized personnel may override AI-generated outputs where appropriate.
Examples include:
Correcting inaccuracies
Adjusting recommendations
Rejecting unsafe outputs
Replacing AI recommendations with human recommendations
Escalating concerns for additional review
Human decisions always take precedence over AI-generated outputs. AI systems are not permitted to override human decisions.
7. WHO PROVIDES HUMAN OVERSIGHT?
Depending on the nature of the AI output, oversight may be provided by:
Licensed Physicians
Medical Specialists
Clinical Reviewers
Health Coaches
Medical, Health & Wellness Board Members
AI Governance Committee Members
Compliance Personnel
Technical Specialists
Reviewers are selected based on the subject matter, risk level, and intended use of the AI output.
8. HIGH-RISK AI ACTIVITIES REQUIRING HUMAN REVIEW
Human review is generally required for:
Diagnostic support outputs
Treatment recommendations
Medication-related recommendations
Risk stratification outputs
Emergency condition identification
Mental health crisis indicators
Clinical decision support outputs
Significant health alerts
Escalation recommendations
Personalized healthcare recommendations
These outputs may not be acted upon without appropriate human oversight.
9. ESCALATION PROCESS
Where concerns are identified, AI outputs may be escalated for additional review.
Examples include:
Potential patient safety risks
Unclear recommendations
Conflicting outputs
Suspected AI errors
Ethical concerns
Bias concerns
Privacy concerns
Regulatory concerns
Escalation pathways may involve:
Senior clinicians
Medical specialists
Chief Medical Officer
AI Governance Committee
Executive Management
10. HUMAN OVERSIGHT OF WEARABLE DEVICE INSIGHTS
Health insights generated through wearable technologies may involve AI-assisted analysis.
Where clinically significant findings are identified:
Additional review may occur
Escalation procedures may be initiated
Human assessment may be recommended
Wearable device outputs should not be considered standalone medical diagnoses.
11. HUMAN OVERSIGHT OF HEALTH COACHING
AI-assisted health coaching services may generate:
Wellness recommendations
Lifestyle suggestions
Educational content
Goal-tracking guidance
Where appropriate:
Human coaches may review recommendations
Recommendations may be adjusted
Additional support may be provided
Health coaching services do not replace professional medical care.
12. HUMAN OVERSIGHT OF AI INCIDENTS
AverCare maintains procedures for reviewing AI-related concerns.
Examples include:
Harmful outputs
Incorrect recommendations
Safety concerns
Hallucinations
Bias concerns
Privacy incidents
Security incidents
Human reviewers investigate incidents and determine appropriate corrective actions.
13. HUMAN OVERSIGHT OF AI MODEL CHANGES
Human oversight applies throughout the AI lifecycle.
Before deployment of significant AI model changes:
Validation activities are performed
Clinical reviews may be conducted
Risk assessments are completed
Governance approvals are obtained
This helps ensure that updates remain safe and appropriate.
14. HUMAN ACCOUNTABILITY
Artificial Intelligence systems are tools.
Responsibility for decisions affecting users remains with:
AverCare
Healthcare professionals
Authorized reviewers
Governance personnel
AI systems are not considered decision-makers and cannot be held accountable for healthcare decisions.
Human accountability remains essential.
15. USER RIGHTS
Users may have the ability to:
Request information regarding AI-assisted processing
AverCare may update this Human Oversight Information Notice periodically to reflect:
Regulatory developments
Governance improvements
Technology updates
Operational changes
Updated versions will be made available through AverCare platforms.
19. VERSION CONTROL
09 June 2026 Date: ______________________________ 09 June 2026
This is the published version of AC-PUB-HOI-162, version 1.0. The signed original is retained on record and is available to customers, auditors and regulators on request via ai-governance@avercare.global.
