Examine the potential and the risks of using Artificial Intelligence in public healthcare delivery, citing recent Indian initiatives.

India's National Telemedicine Service, eSanjeevani, crossed 50 crore teleconsultations in September 2026. It now includes an AI-powered Clinical Decision Support System (AI-CDSS) [1]. AI can help a small number of doctors serve a huge primary-care network. Without safeguards, however, it can also bring new risks of error, bias and exclusion.

Potential of AI in public healthcare

• Decision support at scale: AI-CDSS makes clinical decisions more consistent across 1.42 lakh+ Ayushman Arogya Mandir spokes. It won the Gold Award, National Awards for e-Governance 2026 [1].
• Continuum of care: the AI-supported platform now links to specialist OPDs at AIIMS Jodhpur, Jammu and Patna and to ERSS-112 emergency response [1].
• Inclusive reach: 57.14% of users are women and 16.25% are aged 60+. It works in 15 languages [1].
• Health-worker performance: WHO lists health-worker decision support as a priority digital intervention. It finds that doctor-to-doctor telemedicine may improve how health workers perform [2].

Risks

• Unproven clinical safety: an e-governance award judges service delivery, not medicine. No public data shows how accurate the AI is or how often doctors overrule it [1].
• Accountability gap: the Telemedicine Practice Guidelines (25 March 2020) came before AI-CDSS [3]. If an AI suggestion leads to a wrong prescription, it is unclear who is liable: the doctor, the State or C-DAC.
• Algorithmic bias and digital divide: a model that learns from existing users may work less well for groups that use the service least. WHO warns that digital approaches can widen health inequities [4].
• Infrastructure dependence: the Parliamentary Standing Committee on Health asked for rural broadband in "mission mode" at every health centre [5].
• Data privacy: sensitive health data must be handled in line with the Digital Personal Data Protection Act, 2023.

Way forward

• Independent validation of AI suggestions, plus published outcome data such as referrals completed.
• Update the 2020 Guidelines: the doctor makes the final decision and is accountable for it.
• Digital training for ASHAs and sustained IEC campaigns, as the Committee recommended [5].

Overall, eSanjeevani shows that India's own digital public infrastructure can bring AI-assisted care to the last mile. To deliver on that promise, AI must stay a validated, human-in-the-loop tool and must include everyone. That would bring India closer to the right to health under Article 21 and to SDG 3's universal health coverage.