As India hosts a world AI Impression Summit, rhetoric concerning the transformative potential of AI in healthcare seems to be overshadowing necessary realities and considerations. A special form of dialog on AI came about in Delhi on February 7 throughout the Nationwide Session on Individuals-led AI in Well being, highlighting different approaches based mostly on well being rights, sufferers, and healthcare suppliers. On this dialog, clinicians, public well being specialists, AI engineers, healthcare professionals, and affected person advocates got here collectively to not have a good time AI, however to query it whereas exploring different approaches. Whereas AI is projected as a serious answer to India’s well being challenges, this reflective session goes past the worldwide hype and acknowledges that deploying AI via centralized, commercially pushed programs that override the rights of sufferers, communities and healthcare employees may very well do extra hurt than good.
Globally, the usage of AI in healthcare has proven some promise in particular areas corresponding to picture recognition in radiology, evaluation to assist analysis in managed environments, and workflow help. Nevertheless, systematic critiques have repeatedly proven that instruments that seemingly work nicely in pilot settings don’t carry out nicely in real-world conditions. AI is nice at recognizing and matching patterns, however healthcare includes extra than simply sample recognition, because it includes complicated medical and moral judgments, affected person socialization with explanations and reassurance, and direct bodily care. All of this includes relationships in addition to algorithms.
important to guard rights
The Delhi talks raised sharp considerations about digital extractivism. Who owns well being information, who advantages from derived intelligence, and who bears the dangers? Sufferers have to be defined and empowered, not simply handled as an information supply. Moreover, if AI instruments are educated totally on digitally city populations, caste, gender, regional, and socio-economic biases might grow to be entrenched. Subsequently, the usage of AI in healthcare should be anchored in a powerful rights-based framework. This consists of the best to understanding as a result of sufferers and other people shouldn’t simply have entry to their well being information. They need to have the ability to perceive that. AI programs should remodel complicated medical info into clear and related explanations that assist knowledgeable decision-making. The appropriate to native processing implies that delicate well being information must be processed regionally at any time when attainable by default, somewhat than being centrally managed on company or state-controlled servers. Cloud shares should be specific and revocable. The appropriate to ongoing management implies that consent shouldn’t be a one-time formality. People want to have the ability to withdraw entry to their information, and so they have to be in management not solely of their information but in addition of the insights which are generated from it. The appropriate to fairness and entry implies that AI programs should be audited for bias, accessible throughout geographies and languages, and transparently managed to scale back and keep away from exacerbating well being inequalities. AI-assisted providers developed utilizing public sources must be made freely accessible to be used inside public well being programs. Non-excludability should be assured. Nobody must be denied care as a result of they aren’t participating with an AI system. Non-AI pathways in healthcare should stay accessible and viable.
Complementing human care
The essential precept is that AI ought to complement, not change, human care. AI might assist documentation and information interpretation, however medical choices should be made by accountable human suppliers. Conserving in thoughts that healthcare employees and specialists are the spine of care, people should all the time have AI-assisted capabilities accessible. In a well being system already characterised by precarious working circumstances, there’s a actual threat that AI will justify attrition, casualization, elevated workloads, and algorithmic surveillance of ASHAs and different frontline employees. Approval of AI instruments ought to due to this fact require labor affect assessments, guaranteeing accountability to frontline employees, and specific ensures towards headcount reductions. Any technological progress should improve the competence and dignity of well being employees, not displace them.
The political financial system of AI
The elemental query shouldn’t be whether or not AI will assist, however who it’s going to assist. Present use of AI shouldn’t be impartial. It’s largely constructed right into a monopolistic, profit-driven mannequin. If AI is deployed via industrial platforms that centralize affected person information, it dangers turning into extra corporatized, creating an elite class of care, and getting used for high-cost market growth somewhat than affordable entry. When public information and public funds construct AI programs, their main responsibility is to reinforce public provision, not subsidize company pursuits.
Using AI in India should be based mostly on a well being programs method. AI could be correctly deployed to reinforce main and preventive care and empower sufferers by supporting rational drug use, enhancing referral programs, demystifying hospital billing, and simplifying medical info for customers. Nevertheless, it should be remembered that the challenges for India’s healthcare system should not primarily technical. They’re political, financial, and structural, and embrace continual underinvestment in public well being, lack of educated personnel, insufficient regulation of economic well being care, and excessive out-of-pocket prices. These are institutional failures that can’t be mounted by algorithms.
In conclusion, we should always not anticipate know-how to supply options to essentially coverage or systemic issues (referred to as “techno-solutionism”). As with different applied sciences, AI should contribute to affected person rights, well being fairness, and public targets whereas healthcare employees and specialists stay the spine of care. Healthcare information should belong above all to sufferers and other people, and derived intelligence should be accountable to them. AI and varied applied sciences may also help in shaping the way forward for healthcare in India, however folks, healthcare employees, and public well being should stay firmly on the heart.
Dr Abhay Shukla is a public well being doctor and nationwide co-convener of Jan Swasthya Abhiyan. The views expressed are private. He wish to thank Surajit Nundy, the RAXA crew, and the individuals of the Individuals-Led AI in Well being session for offering worthwhile concepts that knowledgeable this text.
issued – February 20, 2026 12:34 AM IST
