"AI can augment but not replace clinical judgement in healthcare." Discuss with reference to India's regulatory approach to AI-enabled medical devices.

Artificial intelligence now assists diagnosis, imaging analysis and treatment workflows, yet accountability for a clinical decision remains legally and ethically indivisible from the treating doctor. India's response reflects this: AI tools are regulated as devices, while the clinician stays the decision-maker.

How AI augments clinical practice

• Diagnostic support and scale: AI-enabled devices automate stages of screening and reporting, extending scarce specialist capacity to underserved districts.
• Validation infrastructure: BODH, a privacy-preserving benchmarking platform built by IIT Kanpur with the National Health Authority, tests AI models on real-world health data without sharing datasets [1].
• Policy backing: SAHI (Strategy for AI in Healthcare in India), launched with BODH on 17 February 2026 at the India AI Impact Summit, guides governance, data stewardship, validation and monitoring [1].

Why it cannot replace clinical judgement

• Automation bias: WHO warns that large multi-modal models may generate false or biased output and induce over-reliance among providers [2].
• Accountability gap: WHO's principles of human autonomy and traceable responsibility require a human answerable for outcomes — an algorithm cannot be sued or struck off [3].
• Equity and context: models trained on narrow datasets misread rural, paediatric or comorbid presentations that experience detects.

India's regulatory approach

• AI-enabled medical devices are regulated by CDSCO under the Medical Devices Rules, 2017, framed under the Drugs and Cosmetics Act, 1940 — there is no standalone AI law [4].
• Applicants must file an Essential Principles Checklist, verification-and-validation documentation including software, and software version release certificates [4].
• This is thus dual-track: statutory device-level clearance plus a non-binding ethical strategy — regulating the tool, not the decision.

The regulatory design confirms the premise: India licenses AI as an aid subject to safety and performance proof, while clinical responsibility stays with the physician. Going forward, embedding AI literacy in medical curricula, mandating post-market performance surveillance, and issuing binding liability norms under SAHI would let India harness AI's reach while upholding the patient's right to a human healer under Article 21.