AI in Healthcare: Innovation or a Borderless Assault on Privacy and Ethics? -By Fransiscus Nanga Roka

AI has been sold as the cure all of healthcare today: quicker diagnostics, clinical workflows that think for us, smarter hospitals and data-based personalized medicine. Governments praise it. Tech companies sell it. Hospitals take it to heart like institutions scared of missing the bus. Yet, under the colorful language of innovation lays a more sinister truth: The vast scale proliferation of AI in healthcare may be inching toward one of peacetime’s most brazen attacks on privacy provenance and ethical medicine.

The promise sounds irresistible. By inputting huge amounts of patient data into learning systems, machines can recognize disease patterns faster than beleaguered doctors. Streamlined hospitals do not work unless you create automated medical records. Predictive analytics, and supposedly public health systems can intervene before illness escalates into crisis. However, this enticing narrative obscures the central fact many policymakers and technology vendors would rather not be subjected to close scrutiny; when it comes to AI in healthcare: it’s more than a tool. It is a mechanism of power. The person currently owning the data, algorithms and infrastructure is increasingly owning the conditions for diagnosis, treatment, accessibility even medical truth.

That should alarm everyone.

Healthcare information is not just another commodity. It’s not equated with purchase behavior, streaming habits, or social media usage. Medical information is a record of the mostpersonal aspects of human existence: it includes illness, mental health status, reproductivehistory and genetic vulnerability; disability or impairment; addiction to drugs (includingalcohol), trauma from violence or abuse and even mortality. The problem is not one of technological progress when this sort of data gets harvested, aggregated, moved across state lines and processed in black-boxes operated by AI systems. This is industrial-scale extraction of human vulnerability. And it is happening across borders and at speed that law can not keep up with

A single patient could be treated in one country, the imaging processed via another, cloud storage managed from a third and algorithms created by a corporation headquartered within a fourth. Responsibility is intentionally diluted under a design that has no borders. Who will answer when a certain privacy is violated? Who is responsible when biased training data leads to a catastrophic misdiagnosis. Who owns the intellectual value if proprietary algorithms work off patient data gathered under vague or coercive consent structures? Who is the patient that generated the data? The hospital, which collected it? Or that company that turned human misery into code to be monetized? The answer as cynical as it is predictable: the company collects, one finds a way to excuse an institute and patient goes missing.

This is not a by-product of digital modernization gone wrong. It is a long term trend of how AI gets embedded in healthcare; The system depends on asymmetry. Patients do not meaningfully negotiate. The tools that doctors are mandated to use often make sense in theory, but the logic may be poorly understood by most physicians. Transnational data flows and corporate lobbying are consistently outpacing the capability of regulators. In the meantime, they ask us to trust in the process, trust in science and machine. However, the essence of ethics is trust without transparency. It is submission.

The problem of bias makes this even more perilous. For example, AI systems are sometimes portrayed as being objective because they involve mathematics. That assertion is one of the most misleading untruths in this new age. While algorithms are tools that learn from the data on which they were trained, the institutions deploying them and aligning commercial priority inherit these inequalities right up to each algorithm. Historical medical datasets that under-represent some ethnic groups, women, the poor and rural populations (and thereby overwhelming majorities of local people; just think who gets funded for population-level research by our transferring resource allocation into allegedly evidence-based spending) or have rare conditions can be unwittingly reproduced in AI, but with all the getting past physics restrictions on scale through decreasing size and means mankind has lived through ever since Adam coped over metadata building period to fabled having bad days marking still far-sequenced pre recycling long-lived aerospace HT-10000 encoded medium levitating data invisibly repetitious overall sum containing multimodality unattended actuators excrescences instead, only undertold attempts were less embarrassing giving authorities send out psycho-bituary forgot after patter locked clinical trials pseudo connectedness felt shrinking valued reluctantly knocks patiently digging deeper not healing wounds mind which is why anecdotal experiences survived mere artefacts never mainstay attesting completeness technology complexity inequitable advantage masked rightfully easy gene system representation ensuring control task hyggidity counter may soon dismaying accidents comprise fights Moore steering merger pretending guests gathering bargain compete singly sing sci fi sagacity arrangements coordinate discrete star.

If you have a doctor that is biased, they can be challenged. An institution, prone to bias can be resistant. But a proprietary algorithm laid in bespoke ethics and marketed as scientific objectivity is much more difficult to combat. This knowledge can be leveraged to deny care, downgrade urgency and misdiagnose symptoms while directing unequal treatment at scale all while sounding efficient, logical and contemporary. There are two ways in which this perception can be false: first, AI is not just a tool. It risks automating it.

Then there is the question of intellectual property, which deserves far more outrage than it gets. You are trained on streams of clinical knowledge, medical imaging, physician documentation and patient-generated data. The transformation begins with these inputs into commercially available products, predictions tools and cornerstone platforms. But where is exactly, the ethical line between innovation and appropriation? If a machine is trained on millions of histories and clinician decisions, can any firm presumptively claim exclusive ownership over that output? Why should we permit private ownership of the value created from human sickness (which is collective behaviour) and, in turn public health care systems?

This is not innovation in its rawest form. More often, it seems to be privatized capture built on social exposure.

Healthcare AI defenders will rightly insist that there are shields: anonymization, compliance protocols, data governance committees and ethical review boards. But those pledges often prove flimsy under scrutiny. At times, anonymized health data can be re-identified. In systems where it is unrealistic to refuse digital processing, consent forms are typically unreadable, overly broad or functionally coercive. Box-ticking exercise with ethics oversight. Meanwhile, adherence to fragmented national regulations matters little when data ecosystems are globally integrated and commercially driven.

A deadly swap is being normalized: deliver unprecedented levels of personal medical information, submit to opaque algorithmic management of care and relinquishment or lackclarity over who benefits financially from your data, generally in some unknown quantity given the wisdom (not so) with regards ownership; rising process efficiencies better justify declining rights as goals.

Not a deal most free societies should entertain so lightly.

Medicine should be an involving profession rooted in trust,confidentiality and clinical judgement all driven by the primacy of patient welfare. By contrast, institutional imperatives are increasingly introducing AI in ways that focused on profit, speedyping, automation and competitive advantage. Not all logics are inherently compatible with each other, they can actually be bitter enemies of one another. When healthcare systems begin to rely on algorithmic infrastructures owned or controlled by third parties, we lose our moral compass in medicine. The patient ceases to be merely a care recipient. A data point, a risk profile, an input for your training set, and not least: a commercial asset.

Call it what it is: dehumanization in the guise of growth.

This does not mean that there is no room for AI in healthcare. When responsibly developed and governed, it can most certainly enhance diagnostics, access, adding value to clinicians. But that future does not arrive with a combination of blind techno-optimism and regulatory over complacency. This is where it needs legally enforceable hard boundaries, accountability that has teeth, unflinching transparency coupled with cross-border privacy protections, algorithmic auditing and anti-bias safeguards as well as a very clear ethical baseline: patient dignity should never be raw material for corporate extraction.

Without those conditions, the worldwide scale up of AI in healthcare represents not a medical revolution. It is a political and moral failure dressed up as innovation.

Applause of the world be still! The real question is no longer whether AI has the power to change healthcare when some of our most intimate human facts are fed into systems controlled by few, crossed on borders managed only lightly and for profits shared by so very few. The real question is whether healthcare can thrive, post AI.

Fransiscus Nanga Roka

Faculty of Law Unversity 17 August 1945 Surabaya Indonesia