How AI in Healthcare are Similar and Different in Three Continents

 

“Europe asks whether AI is safe. India asks whether it is available. Puerto Rico asks whether it is for us.”    
                            Fable 5.1
 

Three weeks, three continents, one question: what does artificial intelligence become when it lands in any health system around the world? Europe treats AI as something to be governed before it is used. India treats it as something to be deployed before it is perfected. Puerto Rico must use it under rules written for a mainland it does not resemble. The same model is a liability, a leapfrog, or a lifeline, depending on where the AI lands. Here are some of my observations during my three-week journey around the world:

 

 Europe: governance first
Europe's strengths are universal coverage, rich longitudinal data, world-class regulatory and academic depth, and sovereignty pursued as strategy. Its weaknesses are fragmented systems and languages, slow procurement, legacy IT, and a risk aversion that outlasts the risk. Its opportunity is to become the global reference for trustworthy AI, with the European Health Data Space as a multi-national training commons and governance itself as an export. Its threat is that regulation without deployment cedes the field to American and Chinese platforms, and talent to less regulated markets.

Europe's defining feature is that regulation arrived before scale. The AI Act classifies clinical AI as high-risk, layering conformity assessment on top of the Medical Device Regulation and GDPR. A single hospital deploying a triage model may answer to three regimes at once. The result is the world's most deliberate AI, and often its slowest. One speaker at a recent AI in healthcare meeting quipped “Europe is a regulation giant and an innovation dwarf.”

The compensating strength is data. Nordic registries, the NHS's population coverage, and the European Health Data Space promise longitudinal, multi-national training sets no private American system can match. Universal coverage means the incentive question is public value, not reimbursement code. And the continent's insistence on sovereignty: European clouds, European models, European audits is a strategy in a world where health data has become geopolitics. The weakness is fragmentation. Twenty-seven member states, dozens of languages, and hospital IT that predates the euro make Europe a set of pilots searching for a platform.

 

 India: scale first
India's strengths are digital public infrastructure at population scale, engineering talent, low cost, and the political will to leapfrog. Its weaknesses are uneven data quality, paper-bound public hospitals, thin regulatory capacity, and access that varies sharply by geography and income. Its opportunity is frugal AI for the Global South, screening and triage where no clinician exists, and multilingual models for twenty-two official languages. Its threat is bias and harm at scale, privacy erosion, and automation that replaces rather than augments scarce clinicians.

India's defining feature is the opposite: scale arrived before governance. With roughly 1.5 billion people and fewer than one physician per thousand in many districts, the question is not whether AI should replace clinician judgment but whether it can reach patients who have never had any clinician access. Diabetic retinopathy screening in a rural camp is not a workflow improvement in India; it is the first eye examination of any kind in the village's history.

The strength is digital public infrastructure. Aadhaar identity, UPI payments, and the Ayushman Bharat Digital Mission form rails on which health data can move at national scale, something Europe is still negotiating and the United States never built. Add a vast engineering workforce, low cost, and a young population unafraid of a phone-first health system, and India becomes the largest natural laboratory for frugal AI. The weakness is the gap between the rails and the wards. Public hospitals remain paper-heavy, data quality is uneven, and regulation (a 2023 data protection law, CDSCO device rules, evolving guidance from ICMR) is a work in progress. India's risk is the reverse-centaur: clinicians serving the algorithm rather than supervised by it, because supervision costs time no one has.

 

 Puerto Rico: American rules, island realities
Puerto Rico's strengths are US regulatory legitimacy, a bilingual and under-represented population, a pharmaceutical manufacturing base, and a strong diaspora. Its weaknesses are small scale, capped federal funding, physician exodus, a fragile grid, and the absence of an anchor system to lead. Its opportunity is to be the proving ground for Spanish-language and Hispanic-population AI, a hub linking US standards to Caribbean and Latin American needs, and a model of resilience-first health IT. Its threat is the uncritical import of models validated elsewhere, permanent dependency on mainland vendors, and climate shocks that erase progress.

Puerto Rico's defining feature is dissonance. It lives under the FDA, HIPAA, and the CMS reimbursement machinery of the wealthiest health system on earth, yet its Medicaid is capped, its physicians have been leaving for the mainland for a decade, and Hurricane Maria showed that an electronic health record is only as intelligent as the grid beneath it. Its rules are Washington’s but its resources are closer to the Caribbean's.

The strength is unexpected: Puerto Rico is a bilingual, culturally Hispanic population inside a US regulatory perimeter. That makes it an ideal proving ground for AI that must work in Spanish, on a genetically and socially distinct population, under American validation standards (exactly the population most mainland models were never trained on). Its academic medicine, strong pharmaceutical manufacturing base, and diaspora ties give it more technical depth than its size suggests. The weakness is scale and capital. There is no single health system large enough to buy, build, or govern AI on its own, and the brain drain removes the very people who would lead it.

 

 What is the same, what is different
The similarities are more striking than the distances suggest. All three face clinician shortages that AI is asked to relieve. All three are importing foundation models trained elsewhere on populations that are not theirs. All three have discovered that the hard part is not the model but the workflow, the data plumbing, and the trust of the clinicians expected to use it. And in all three, I feel that the ardent enthusiasm of the conference ballrooms run well ahead of the reality of the ward.

The differences are interesting to observe. In governing posture, Europe regulates and then deploys, India deploys and then regulates, and Puerto Rico is governed from elsewhere. In the scarcity each is solving, Europe seeks cost and quality, India seeks access itself, and Puerto Rico seeks workforce and capital. In data, Europe holds deep national registries, India holds national digital rails, and Puerto Rico holds a distinct and underrepresented population. In pace, Europe is deliberate, India explosive, and Puerto Rico constrained. And in chief risk, Europe fears paralysis, India fears unsupervised automation, and Puerto Rico fears dependency and exclusion.

What each venue can teach the others
Europe can teach India that trust, once lost at scale, is not recoverable at scale. India can teach Europe that a perfect framework protecting no one is not ethics; it is delay. Puerto Rico can teach both that the true test of a medical AI is not its performance on the population that built it but on the population that was never asked.
My humble conclusion is a simple one. AI in healthcare is not one field with regional flavors, it is rather three different dimensions on the same technology: safety, access, and belonging. The regions that win will be the ones that learn to strategically place all three.

 

My very special thanks to Dr. Amit Dey of Delhi, India and Dr. Orvil Martinez of Puerto Rico and their teams and families. I have also said that healthcare AI is forged by human-to-human relationships and esprit de corps; they are supreme proof of this.