Examine the ethical and governance challenges associated with the deployment of AI in clinical decision-making, with reference to WHO guidelines.

WHO's 2021 guidance Ethics and Governance of Artificial Intelligence for Health — the first global framework of its kind — laid down six consensus principles for AI in health [1]. As algorithms move from research labs into diagnosis and prescription, these principles expose deep ethical and regulatory gaps that clinical AI must resolve before scale.

Ethical challenges

• Bias and inequity: models trained on unrepresentative data can misdiagnose under-represented groups; WHO's guidance on large multi-modal models flags generation of misleading or biased outputs as a core risk [2].
• Autonomy and consent: patients often cannot know that an algorithm shaped their diagnosis, undermining informed consent — WHO stresses human autonomy and transparency [1].
• Accountability vacuum: when an opaque "black box" errs, liability between clinician, hospital and developer remains unsettled; WHO's emphasis on responsibility and accountability addresses precisely this [1].
• Equity of access: AI concentrated in urban tertiary hospitals risks widening the digital divide rather than closing the access gap [2].

Governance challenges

• Regulatory lag: WHO's LMM guidance issues 40+ recommendations urging governments to enact laws and create assessment bodies — evidence that regulation trails deployment [2].
• Validation deficit: without independent testing on local data, accuracy claims stay unverified. India's BODH platform, built by IIT Kanpur with the National Health Authority, seeks privacy-preserving benchmarking of health AI models [3][4].
• Data stewardship: aggregating records for training strains privacy safeguards, demanding robust consent-management architecture under ABDM [4].
• Capacity gap: regulators and clinicians lack the technical skill to audit algorithms.

India's SAHI (Strategy for AI in Healthcare for India), launched in 2026, is significant precisely because it frames itself as a governance framework rather than a technology push [3]. Aligning such strategies with WHO principles — mandatory bias audits, explainability standards and clinician-in-the-loop safeguards — can make AI a trusted partner in realising the right to health under Article 21.