AI Disclosure
How artificial intelligence is used within the Ukuqala Platform
All AI-generated content is labelled clearly and requires physician review before any clinical use. AI does not diagnose, prescribe, or communicate directly with patients as a healthcare provider.
Overview
Ukuqala Health Technologies uses artificial intelligence (“AI”) technology to provide certain features within the Platform. This Disclosure explains which features use AI, how AI works, its limitations, the safeguards in place, and your rights.
We comply with applicable and emerging AI regulations including:
California AB 3030
Healthcare AI Disclosure (effective 1 Jan 2025)
California AB 489
AI Identity Disclosure
Texas HB 149 / TRAIGA
AI Healthcare Disclosure (effective 1 Jan 2026)
EU AI Act
Where applicable to EU users
WHO AI for Health Ethics
Guidance on AI governance (2021)
AU Digital Transformation Strategy
Africa 2020–2030 framework
AI-Powered Features on the Platform
AI Clinical Assistant (Chatbot)
Provider: Mistral AI
What it does
Provides physicians with a conversational AI interface for clinical queries, protocol lookups, drug references, and ICD-10 suggestions.
Limitation
Responses reflect the AI's training data, which may not include the most recent guidelines or local African disease contexts. The physician must independently verify all responses.
Symptom Analysis & RSE Risk Stratification
Provider: RSE Engine + Mistral AI
What it does
Processes structured symptom inputs to generate a preliminary risk stratification score (low/medium/high) and suggested clinical pathway for physician review.
Limitation
Risk stratification is a triage aid only. It does not diagnose. The physician must assess the patient holistically and independently of the risk score.
Clinical Notes Assistance
Provider: Mistral AI
What it does
Assists physicians in structuring SOAP clinical notes by suggesting format and prompting for key fields.
Limitation
AI-suggested note content is a draft only. The physician is responsible for the accuracy and completeness of all finalised notes.
ICD-10 Code Suggestion
Provider: Rule-based engine with AI-enhanced matching
What it does
Suggests relevant ICD-10 diagnostic codes based on clinical keywords or symptoms entered by the physician.
Limitation
Suggested codes must be verified by the physician for clinical accuracy for each specific patient.
Drug Interaction Alerts
Provider: Rule-based interaction database with AI-enhanced alerting
What it does
Alerts physicians to potential drug-drug interactions when adding medications to a prescription.
Limitation
The database may not capture all interactions, particularly with traditional/herbal medicines prevalent in African markets. Physician pharmacological judgment is essential.
What AI Does NOT Do on This Platform
AI does NOT Diagnose patients
Diagnosis is a clinical act performed exclusively by licensed physicians.
AI does NOT Prescribe medication
Only a licensed, verified physician can issue a prescription.
AI does NOT Communicate directly with patients as a physician
Any AI-generated content that reaches a patient does so through the physician's review and decision.
AI does NOT Make treatment decisions
All treatment decisions are made by the physician.
AI does NOT Replace clinical judgment
It augments it. The physician is always in control.
AI does NOT Represent full African disease coverage
Its outputs may not reflect all diseases prevalent in Central, West, East, or Southern Africa.
Mandatory AI Disclosure Label
All AI-generated content within the Platform is labelled with:
AI-Powered — For physician review only
“This content was generated by artificial intelligence and may contain errors. The treating physician must independently verify before clinical use.”
This label appears on:
- Every AI chatbot response
- RSE risk stratification outputs
- AI-assisted clinical note drafts
- ICD-10 code suggestions marked as AI-generated
- All AI-generated drug interaction alerts
AI Provider — Mistral AI
Mistral AI SAS
15 Rue des Halles, 75001 Paris, France
| Data Processing Term | Status |
|---|---|
| Patient-identifying data transmitted to Mistral AI | No — data is anonymised/pseudonymised before AI processing |
| Your data used to train Mistral AI models | No — prohibited under commercial agreement |
| Data infrastructure location | EU-based; cross-border transfer safeguards apply |
| GDPR compliance | Yes — Mistral AI operates under GDPR obligations |
Known Limitations and Risks of AI
| Risk | Description | Mitigation |
|---|---|---|
| Hallucination | AI may generate plausible-sounding but factually incorrect clinical information | Mandatory physician review; 'For review only' labels |
| Training data bias | AI trained predominantly on Western medical literature may underperform for African disease presentations | Physician override always available; local guidelines take precedence |
| Outdated knowledge | AI knowledge cutoff may not reflect the most recent guidelines or drug approvals | Physicians must consult current references; AI is not the primary guideline source |
| Automation bias | Physicians may over-rely on AI suggestions without independent judgment | UI design actively discourages passive AI reliance |
| Misclassification | Risk stratification tools may under- or over-estimate risk in individual patients | RSE scores are triage support only; physician assessment is primary |
Human Oversight — The Physician Is Always in Control
Ukuqala's AI design philosophy is Human-in-the-Loop (HITL):
No AI output is acted upon without physician review and confirmation
Physicians can dismiss, override, or ignore any AI suggestion at any time
AI features can be toggled off in account settings
Finalised clinical documents reflect the physician's decisions, not raw AI outputs
AI suggestions are never automatically sent to patients
AI does not impersonate licensed healthcare professionals
Patient Rights Regarding AI
Right to know
You have the right to ask your physician whether AI tools were used during your consultation.
Right to explanation
You have the right to ask your physician to explain any AI-generated output that influenced your care.
Right to human review
You have the right to request that your physician not use AI tools during your consultation.
Right to object
In jurisdictions where data protection law provides the right to object to automated processing, you may exercise this right by contacting [email protected].
To report a concern about an AI-generated output: [email protected]
This AI Disclosure forms part of the Ukuqala Terms of Service. Prepared in compliance with California AB 3030, AB 489, Texas TRAIGA, the EU AI Act, and WHO AI for Health guidance. Effective 1 May 2026 · Version 1.0