There is a misplaced argument about the use of artificial intelligence (AI) in healthcare that echoes the dystopian fears of Hollywood films like The Terminator—a vision of machines replacing doctors, automating decision-making, and reshaping medicine overnight. This unsettles patients, frustrates clinicians, and sparks debates on AI ethics, particularly in Low- and Middle-Income Countries (LMICs) or the Global South, such as India, which already faces disproportionately fewer health resources. Yet, while we fixate on these exaggerated fears, a far more pressing threat is going unnoticed: who controls AI, who profits from it, and what happens to the data that powers it?
The answer is troubling—India is sleepwalking into a new form of digital colonisation. Private health data is being extracted at scale by foreign corporations, converted into proprietary models or intellectual property (IP), and eventually sold back at a premium. This trend mirrors the exploitative practices of the pharmaceutical industry, where global corporations historically tested drugs on the world’s poorest populations without fair compensation or adequate safety measures—a phenomenon well-documented in Sonia Shah’s The Body Hunters. Those who provide the resources—whether patient data or human subjects—rarely see the benefits of the innovations they help create, blatantly disregarding international ethical guidelines such as the Helsinki Declaration.
India’s overlapping vulnerabilities—including grossly disproportionate burden (60%) of non-communicable diseases (NCDs) born by the growing poor population and weak personal data regulations—makes it a soft target for healthcare corporations from the Global North to harvest sensitive health data. Furthermore, this predatory exploitation leverages systemic inequities (poverty, disease, and governance gaps) to gain geopolitical control over personal and anatomic data, from X-rays to genetic profiles, exposing a much deeper intersectional crisis. For India to secure its position in AI-driven healthcare and improve public health outcomes, it must invest in domestic innovation and data sovereignty rather than adopt policies that serve corporate profits.
Who controls innovation?
Over the last 10 years, corporate tax revenues have declined, creating an annual deficit of ₹1.5–2 lakh crore, with rates slashed from 30% to 22%. What does this have to do with healthcare? Everything! By prioritising capital accumulation over domestic investment, India’s economic policies have left essential services like healthcare chronically underfunded. The latest National Health Account (2021–22) reveals a grotesque imbalance: households shoulder 45.11% of total healthcare spending through out-of-pocket expenditure (OOPE), while the central government’s contribution languishes at less than 3% of the GDP.
While the comparatively wealthy Indian upper class bears these expenses with private health insurance (8.48% of OOPE), the bottom 80%, already crushed by stagnant wages and rising living costs, face a cruel reality: subsidising corporate profits through taxes while being denied basic care. Also, in a country where newborns die due to delayed payments for medical oxygen supplies, the notion of AI improving public health outcomes risks appearing not just distant but dystopian without a fundamental shift in priorities.
This innovation-investment paradox in India and similar LMICs has become a hotbed for Western trade surpluses funnelled into AI research. Over the last decade, over £68 billion has been invested globally in healthcare AI. While AI models in-theory converge in performance over time, researchers are forced to seek diverse, high-volume datasets to retain a competitive edge—spurring relentless demand for health data from LMICs such as India. However, the returns from patented algorithms and precision therapies trained on Indian bodies flow westward, priced for New York, not New Delhi. Until India recalibrates its priorities from regressive levies to equitable innovation, its AI ambitions will remain a mirage. True progress demands more than algorithms; it requires a reckoning with who controls them and whom they serve.
Data harvesting: The new colonial commodity
AI research is inherently data-intensive. In healthcare, this poses another paradox—medical data is among the most sensitive personal information, ideally owned and controlled by individuals, making its access highly restricted. Yet, in 2023, over 5,000 peer-reviewed AI healthcare papers were published—marking a 1.5x rise from 2022—defying these constraints. How does this happen? Through the unrestricted cross-border flow of healthcare data from the Global South.
Strict regulations make healthcare data from developed Western economies difficult for global corporations to access, often resulting in ‘techlash’ when they attempt to bypass them. Meanwhile, the Global South, keen on receiving Western funding via research grants to bridge budget deficits, often trades restricted patient data. This ‘lawful’ exchange is achieved through de-identification or anonymisation, where personal identifiers (names, addresses) are stripped from patient records to create anonymised datasets. Global corporations tactically fund clinical trials through partnerships with low-cost, charitable healthcare providers in LMICs. Blanket consent—extracted by exploiting power imbalances between doctors and patients—is rubber-stamped by under-resourced Institutional Review Boards (IRBs).
While studies are published to fulfil contractual obligations, the coded, structured datasets migrate westward. Such practices have historic precedence: For decades, unethical trials—like the US Centers for Disease Control and Prevention (CDC)-funded placebo-controlled study for the HIV drug tenofovir in Thailand, which withheld treatment from participants—have been offshored to circumvent US Food and Drug Administration ethics boards. Today, the same extractive logic applies to healthcare data: patient records and genetic information are the new colonial commodities. The result? LMICs like India remain a raw-data supplier in the global AI supply chain, with its healthcare disparities and disease burden unaddressed, while corporations profit from innovations trained on the bodies of its poor.
Breaking free: Policy imperatives
Our health is not a choice but a calculus of power—where you live determines if your body will be healed or harvested. The Global South’s disproportionate tuberculosis, HIV, and cholera burden stems not from epidemiological chance but from centuries of engineered vulnerability—a legacy now extending into the digital age through predatory data extraction. Today, chronic underinvestment in healthcare infrastructure, intersecting with predatory data harvesting, entrenches a perverse duality: populations battling preventable diseases simultaneously serve as unwitting subjects for medical innovation patented in the Global North.
Nowhere is this modern colonialism clearer than in India's healthcare data—a trove of genetic profiles, diagnostic scans and treatment histories being strip-mined by foreign corporations. While algorithms trained on Indian bodies generate billions in proprietary medical AI, the Union Budget's token ₹50 crore (£5 million) for domestic research lays bare the dialectical contradiction: we are funding our own obsolescence. Individual health behaviours—diet, exercise—are rendered irrelevant against social determinants such as inadequate sanitation and unaffordable treatments. Breaking this cycle demands reclaiming sovereignty over the 21st century's most valuable resource—our biological data.
Treat data as a national resource
Like oil or minerals, health data must be governed by principles of resource nationalism. This requires:
Trusted Research Environments (TREs): Mandate that Indian patient data remains onshore, accessible only to domestic researchers and vetted AI developers. TREs should operate as digital fortresses, preventing cross-border data plunder while fostering collaboration under strict public oversight.
Data commons with teeth: Build a national health data repository where anonymised datasets are pooled for public-interest AI projects. Private firms can access these commons only if they reinvest profits into India’s healthcare infrastructure.
2. Legislate sovereignty, not symbolism
Current ‘AI for All’ slogans ring hollow without binding laws to back them. India needs:
Sovereign AI regulations: Any AI model trained on Indian health data must first address India’s public health priorities—reducing maternal mortality, diabetes blindness, or TB detection—before being monetised abroad.
Benefit-sharing mandates: Legally require foreign corporations to share foreground IP rights or royalties from AI tools derived from Indian data. For example, a cardiac algorithm trained on AIIMS’ ECG scans should grant India a share in the IP or provisions for using the technology to benefit patients.
Invest in homegrown technology
In the event of financial deficits, prioritise scaling successful AI-powered models like Kerala’s:
Scale Kerala’s model nationwide: Kerala’s AI-driven diabetic retinopathy programme ’Nayanaritham’, funded domestically and tailored to local needs, proves innovation need not be outsourced. Replicate this with state-led missions targeting India’s top disease burdens.
4. Embed Patient and Public Involvement (PPI) in Innovation:
To ensure innovation’s alignment with community needs and ethical imperatives such as patient rights, India must institutionalise Patient and Public Involvement (PPI) frameworks in AI development
o Co-design with communities: Mandate PPI groups in R&D—comprising patients, caregivers, and grassroots health workers—to partner in designing, testing, and auditing AI tools similar to those in the west.
o Transparency and trust: PPI collectives should co-govern access to localised datasets in TREs, ensuring AI-development priorities address regional interests rather than corporate interests.
o Ethical guardrails: Empower PPI groups to veto exploitative data practices, such as blanket consent forms, and negotiate benefit-sharing agreements.
The alternative is permanent servitude in the digital plantation, where our bodies remain the soil, others reap the harvest, and geography still dictates who lives and who dies. The current budgetary allocation for AI research is woefully inadequate and risks undermining the very population whose anonymised health data fuels global AI innovation. True progress demands treating healthcare data as a non-negotiable national interest, not a giveaway to the highest bidder. The question is, will it?
Alok Raj Kochuparampil is a Senior Product Manager with over eight years of experience driving digital transformation and AI product strategy in public health settings. He has led pioneering projects with the NHS, focusing on explainable AI and patient-centred innovation. Views expressed are the author's own.