Artificial intelligence is no longer confined to research laboratories—it is increasingly becoming part of everyday healthcare. AI tools can assist clinicians with finding information, interpreting medical data, supporting clinical decisions, communicating with patients, conducting research and reducing routine administrative work.
General-purpose generative AI tools such as ChatGPT, Microsoft Copilot, Google Gemini and Claude can help doctors explain complex medical information in simpler language, draft patient information materials, summarise documents and develop teaching resources. However, their outputs may be inaccurate or incomplete and must always be independently verified.
Evidence-search tools use AI to help clinicians locate and synthesise medical literature more efficiently. Examples include OpenEvidence, Consensus, Elicit and Perplexity. These tools may be useful when exploring clinical questions, but they should complement—not replace—trusted evidence-based resources and critical appraisal.
More specialised clinical AI tools are also emerging. Some assist with interpreting medical images such as chest X-rays, retinal photographs and mammograms. Clinical decision-support applications may suggest possible diagnoses, identify drug interactions or help clinicians consider appropriate investigations. AI-enabled “ambient scribe” tools can listen to a clinical consultation—with appropriate consent and privacy safeguards—and prepare a draft clinical note for the doctor to review.
The usefulness, cost and availability of these tools vary considerably, particularly in low- and middle-income countries. Clinicians should therefore select tools according to a clear clinical purpose and consider their accuracy, evidence base, privacy protections, affordability and relevance to the local population. AI should be regarded as a supportive tool rather than an independent decision-maker: the clinician remains responsible for verifying its output, applying professional judgement and making the final clinical decision.
WISER Framework
Artificial intelligence can support clinicians, but its value depends on how wisely it is used. The WISER framework offers five practical principles for responsible use:
W — Work with a clear clinical purpose. Use AI to address a specific need rather than simply because the technology is available.
I — Interrogate and verify the output. AI can produce inaccurate, incomplete or fabricated information. Important claims should be checked against reliable clinical guidelines and evidence.
S — Safeguard privacy and confidentiality. Do not enter identifiable patient information into public AI systems unless the tool has appropriate security, consent and institutional approval.
E — Exercise professional judgement and remain accountable. AI may assist decision-making, but it cannot understand the patient’s complete clinical and social context. The clinician remains responsible for the final decision.
R — Review performance, equity and competence. Consider whether the tool performs reliably in the local population, including resource-limited settings, and maintain the skills needed to work safely without it.
Used wisely, AI can strengthen rather than replace clinical practice. The goal is not to hand decisions over to technology, but to combine its capabilities with evidence, ethical practice, professional judgement and compassionate patient care.