Why more information doesn’t always lead to better healthcare decisions.
For years, “Dr Google” has been a familiar part of the healthcare journey. Research suggests that 65% of people search online before contacting a doctor, often arriving at consultations with pre-formed ideas about their symptoms.
Today, that behaviour is evolving, with AI-powered tools increasingly becoming the first point of consultation. Harvard Medical School notes that the use of digital symptom checkers has surged, increasing by as much as 134% over the past year. These platforms no longer simply provide links – they deliver structured responses, including possible diagnoses and suggested next steps, often with a level of confidence that feels authoritative.
In many ways, this shift is making patients more informed and better equipped to engage in meaningful conversations about their health. As Dr Charl van Loggerenberg, General Manager of Emergency Medicine at Life Healthcare, notes: “We are seeing patients arrive with more refined and structured questions, which is a positive shift. The challenge is that those questions are sometimes built on incomplete information.”
However, as AI becomes more accessible and persuasive, it introduces a critical tension: the illusion of certainty in a space that demands clinical nuance. Patients are no longer navigating multiple sources – they are receiving interpreted answers, often without full visibility into their accuracy, completeness, or context. The problem is this: AI can provide convincing information based on what it is given. In healthcare, however, that input is rarely complete, and Dr van Loggerenberg cautions that what remains missing is clinical judgement. “AI does not have the ability to interpret, prioritise and contextualise information in real time.”
At the same time, the way health content is created and surfaced online is evolving. This places a responsibility on healthcare organisations to ensure that credible, clinically sound information is available for AI tools to draw from, an important step in supporting more reliable AI-assisted health journeys.
Where AI falls short: urgency, context and risk
More information does not always lead to better decisions.
In clinical settings, particularly in emergency care, decisions are not based on symptoms alone. They require context and urgency – as two patients may present with the same symptoms but require entirely different responses.
Dr van Loggerenberg recalls two patients who presented late to an emergency unit – one with malaria and the other with tick bite fever. Both conditions can be life-threatening if left untreated. In each case, critical context was missing from the patients’ online searches, including travel history and, in one case, a visible physical sign. The issue was not incorrect information, but incomplete interpretation – a distinction that can carry serious consequences.
Patients may describe symptoms accurately but omit key details that fundamentally change the clinical picture. More importantly, AI lacks the ability to triage – the process of prioritising patients by urgency, grounded in training, experience, and real-time assessment, rather than simply matching symptoms to conditions.
As Dr van Loggerenberg explains: “AI can suggest what a condition might be, but it cannot reliably tell you how urgent it is. That is where clinical judgement is critical.” This creates what he describes as a “triage error” – a gap between actual and perceived urgency. This gap cuts both ways: patients may be falsely reassured and delay care, or become unnecessarily alarmed. Neither outcome serves the patient.
Beyond triage, several risks are emerging: delayed care, where reassurance from AI leads to postponed medical attention; confirmation bias, reinforcing existing beliefs regardless of clinical reality; data privacy concerns, where sensitive health information may be unknowingly exposed; and AI limitations, including confident but incorrect outputs that create false confidence.
Dr Karisha Quarrie, Chief Medical Officer at Life Healthcare, emphasises: “AI must remain assistive, not autonomous. It can enhance care, but it cannot replace clinical responsibility.” In South Africa, this position is supported by frameworks such as the Health Professions Council of South Africa’s (HPCSA) guidelines and the Protection of Personal Information Act (POPIA), which emphasise clinician accountability, informed consent, and the protection of clinical data.
A critical but often overlooked risk is data exposure. Uploading scans or results onto AI platforms may feel harmless, but these contain highly sensitive personal information. In addition, many AI models are trained on international datasets that may not fully reflect South African populations, introducing further risk if used without appropriate validation. As Dr Quarrie cautions: “Patients need to understand that their health data is highly sensitive. If you’re using AI tools, always anonymise your information before sharing it.”
From risk to responsibility: the role of AI in healthcare
Despite these risks, AI has a meaningful role in healthcare.
At Life Healthcare, AI is being explored and progressively integrated within established clinical governance frameworks. The focus is on supporting decision-making, identifying risks earlier, improving efficiency, and enhancing patient communication. This includes clinician-facing tools such as decision support and diagnostic assistance, with the clear intention that AI augments, rather than replaces, clinical expertise.
It also has the potential to support clinicians in high-pressure environments by flagging risks such as drug interactions or potential oversights. In this context, AI serves to strengthen, not substitute, professional clinical expertise and judgement. This reflects a broader commitment to responsible innovation, where technology is thoughtfully adopted to enhance care while maintaining strong clinical oversight.
AI is here to stay – and its potential in healthcare is significant. Used responsibly, it can empower patients and support better outcomes – but its role must be clearly understood. In practice, this means:
- Patients should view AI as a starting point – not a diagnosis – and use these tools responsibly
- Clinicians should use AI as a support tool – not a substitute for judgement
- Healthcare systems must ensure AI is guided by transparency, governance and trust
For healthcare providers such as Life Healthcare, the evolution of AI presents both an opportunity and a responsibility – to guide patients in using these tools safely and effectively, while continuing to build systems that combine innovation with strong clinical oversight, patient education, and clear accountability.
But this responsibility does not sit with healthcare providers alone. The future of healthcare, particularly in the era of AI, is inherently collaborative and requires shared responsibility across patients, clinicians, and healthcare systems to ensure AI is used safely and effectively. Building this future will depend on trust, transparency, education, and robust governance. AI can enhance the healthcare journey. The best outcomes will always come from combining technological capability with clinical judgement, common sense, and a qualified healthcare professional at the centre of care.
