Interoperable health data is key to AI in healthcare, but raises consent and privacy concerns. Discuss.

The Ayushman Bharat Digital Mission has built a health-data backbone around the 14-digit ABHA, three registries and three gateways — HIE-CM, UHI and NHCX [1]. With over 97 crore ABHAs and 121 crore linked health records [2], India now holds the raw material for health AI; the same interoperability, however, makes consent and privacy the decisive safeguards.

Why interoperability is the foundation for AI

• Standardised, machine-readable data: NITI Aayog has stressed that interoperable, standardised records are the base for health research and responsible AI use [2]; without common treatment codes, a shared record can be exchanged but not understood.
• Fraud detection: NHA already uses AI/ML triggers and real-time dashboards on claims — leading to 1,114 hospitals de-empanelled and ₹122 crore in penalties [3].
• Continuity of care and portability: linked records let a migrant worker's history travel with them; NHCX places hospitals and insurers on one claims platform [1].
• Policy targeting: granular data helps attack the residual burden, as out-of-pocket expenditure fell from 62.6% (2014-15) to 39.4% (2021-22) but remains high [4].

Consent and privacy concerns

• Quality of consent: HIE-CM makes sharing consent-based [1], but a patient asked to approve at a hospital counter to receive treatment can rarely refuse meaningfully.
• Irreversible harm: health data is the most sensitive category — a leaked HIV or cancer record cannot be undone; ABDM has therefore layered cyber-security measures [5].
• Legal frame still maturing: the DPDP Act, 2023 supplies purpose limitation and data minimisation [6], but secondary use — insurers risk-profiling from ABHA-linked histories — needs explicit guardrails.
• Exclusion and bias: only 5.6 lakh facilities and about 11 lakh professionals are registered [2], so datasets under-represent small clinics, and AI trained on them may misjudge the unregistered poor.

Interoperability and privacy are complements, not trade-offs: auditable consent artefacts, anonymised research datasets, strict purpose limitation and independent security audits can let AI serve care without making the patient the product. Treated as a public good under a firm DPDP framework, India's health data stack can advance both equity and innovation.