At AverCare, Artificial Intelligence (AI) is designed to support, not replace, the expertise, judgement, and decision-making of qualified healthcare professionals.
We believe that responsible AI requires meaningful human oversight, particularly when AI is used in healthcare environments where recommendations may influence health and wellbeing. Our Human-in-the-Loop (HITL) approach ensures that appropriately qualified individuals remain responsible for reviewing AI-generated information before clinical or patient-related decisions are made.
Human-in-the-Loop (HITL) means that AI systems assist people by providing information, insights, or recommendations, while appropriately qualified healthcare professionals remain responsible for reviewing, validating, and making final decisions where required.
Our AI systems are designed to support healthcare professionals; not to act independently.
Where AI generates information that could influence medical or clinical care, appropriate human review is required before action is taken.
Depending on the nature of the AI output, this review may include licensed physicians or other appropriately qualified healthcare professionals with expertise relevant to the specific clinical area.
AverCare AI does not independently:
Final responsibility for healthcare decisions always remains with qualified healthcare professionals.
Our level of human review is proportionate to the potential impact of the AI output.
These types of information may be subject to periodic quality review.
These outputs may require review before they are applied in a patient-specific context.
These outputs require review by appropriately qualified healthcare professionals before they are relied upon for clinical decision-making.
Our AI systems are designed so that human reviewers can:
Human judgement always takes precedence over AI-generated outputs.
Patient safety is a core principle of our AI governance programme. Where AI identifies information that may indicate a medical emergency, significant clinical concern, or potential patient safety issue, appropriate escalation processes are followed to ensure timely human review.
AI systems are not designed to independently manage emergency situations.
We believe users should understand when AI is being used.
Where appropriate, AverCare informs users that:
judgement.
required.
To maintain the quality and safety of our AI systems, AverCare regularly performs activities such as:
These activities help us continuously improve the reliability, safety, and performance of our AI systems.
Our Human-in-the-Loop approach forms part of AverCare's broader Responsible AI programme, which is built on the principles of:
This is the published version of AC-COM-HITLS-202, version 1.0. Controlled originals are retained on record and are available on request via ai-governance@avercare.global.
