Introduction

Artificial intelligence is no longer a distant concept in medicine. Across Indian hospitals, diagnostic centres, and healthtech platforms, AI-powered tools are already reading radiology scans, flagging drug interactions, predicting patient deterioration, and even suggesting treatment pathways. The speed, accuracy, and scale that AI brings to healthcare are genuinely promising. But these capabilities raise a question that the medical and legal communities in India are only beginning to confront seriously: when an AI system influences a clinical decision, and when a patient is not fully aware of that influence, who holds responsibility for the outcome?

At the heart of this debate lies one of medicine's most foundational ethical principles, namely informed consent. For decades, informed consent has meant that a patient receives clear, honest, and complete information about a proposed treatment or procedure before agreeing to it. The patient is supposed to understand what will be done, why, what the risks are, and what alternatives exist. The responsibility for obtaining that consent rests squarely with the treating physician. But when an algorithm quietly drives part of that clinical reasoning, the traditional framework begins to show its limits.

Understanding the Role of AI in Medical Decision-Making

To understand where responsibility lies, it is important to first understand how AI currently operates inside clinical environments. Most healthcare AI tools in India and globally function as decision support systems rather than as autonomous agents. That means they do not replace doctors but instead provide recommendations, alerts, predictions, or risk stratifications that clinicians are expected to review and act upon.

In practice, however, the influence of these tools can be quite significant:

• An AI tool that flags a chest X-ray as suspicious for malignancy may lead a radiologist to recommend a biopsy.
• A predictive algorithm that assigns a patient a high risk score for sepsis may lead to early aggressive intervention.
• A clinical decision support system that recommends a specific drug dosage may shape a prescription without the patient ever knowing an algorithm was involved.

In each of these scenarios, the AI is not making the final decision in a legal sense, but it is unquestionably shaping the direction of care. And in most cases, patients are not informed that AI was part of the process.

What Informed Consent Actually Requires

In India, the legal and ethical framework for informed consent is shaped by guidelines from the National Medical Commission (NMC), the Indian Council of Medical Research (ICMR), and broader principles laid down in the Indian Medical Council (Professional Conduct, Etiquette and Ethics) Regulations. The foundation of informed consent is that a patient must be given material information, meaning information that a reasonable patient would consider relevant to their decision.

When an AI system plays a meaningful role in diagnosis or treatment planning, a strong argument can be made that this constitutes material information. If a patient knew that a risk prediction model was flagging them for a particular intervention, they might want to ask how accurate that model is, what data it was trained on, whether it was validated on Indian patient populations, and what happens if the algorithm is wrong. These are not unreasonable questions. They are exactly the kind of questions that informed consent is designed to allow patients to ask.

Currently, Indian law does not explicitly mandate disclosure of AI involvement in clinical decisions. The NMC guidelines and the Digital Information Security in Health Act (DISHA), which is still under development, have not yet established a clear standard. This legal ambiguity leaves both patients and providers in an uncertain position.

The Question of Accountability

The accountability question in AI-influenced medicine is genuinely complex because it involves multiple parties, each of whom contributed to the outcome.

The treating physician remains the most visible point of legal accountability. Under current Indian medical law, the doctor who makes or signs off on a clinical decision is the one who bears professional and legal responsibility for that decision. If an AI tool provided a flawed recommendation and the doctor acted on it without adequate verification, the doctor may still be held liable under the Consumer Protection Act or under claims of medical negligence.

The AI developer or technology company that built and sold the tool occupies a different but equally important position. If the algorithm was poorly designed, inadequately validated, or trained on biased data that does not reflect Indian patient demographics, the company that created it carries a share of the moral and potentially legal responsibility. India does not yet have a dedicated AI liability law, but the Information Technology Act and general product liability principles under consumer law may apply in some circumstances.

The hospital or healthcare institution that procures and deploys an AI tool without properly vetting it, training its staff on its limitations, or informing patients of its use also bears institutional responsibility. Adopting AI for cost savings or marketing purposes without ensuring clinical safety and ethical deployment is a governance failure.

Finally, the regulatory ecosystem itself plays a role. The Bureau of Indian Standards and the Central Drugs Standard Control Organisation (CDSCO) have begun examining how AI-based medical devices should be classified and regulated, but comprehensive AI-specific guidance for clinical deployment remains limited.

AI Bias and the Indian Patient Population

One of the most underappreciated concerns in the Indian context is the problem of algorithmic bias. Most advanced clinical AI systems in use today were trained primarily on datasets from Western countries, specifically from patient populations in the United States, Europe, and parts of East Asia. These datasets differ from Indian patient populations in significant ways, including genetic variation, disease prevalence patterns, environmental exposures, nutritional status, and comorbidity profiles.

When an AI model trained predominantly on non-Indian data is applied to patients in India, the risk of misclassification or flawed prediction increases. A cardiovascular risk algorithm calibrated to one population may not perform equally well in a patient from Bihar or Tamil Nadu. A diagnostic imaging tool trained on one type of imaging equipment may behave differently on machines commonly used in tier-2 Indian cities.

This is not merely a technical concern. It is an informed consent concern. If patients are not told that the AI tool being used to guide their care may not have been validated on people like them, they are being denied information that is genuinely relevant to their decision about whether to proceed with a recommended course of action.

What Responsible AI-Informed Consent Should Look Like

While India waits for formal regulatory standards, responsible healthcare providers and institutions can already take meaningful steps toward ethical AI deployment and honest patient communication.

First, disclosure should become standard practice. Before an AI tool is used in a clinically significant decision, the patient should be informed in plain language that an AI system will assist in the analysis, what kind of tool it is, and what role it plays. This does not require turning every consultation into a technology lecture. A brief, clear explanation is sufficient.

Second, physicians should be adequately trained not just in how to use AI tools, but in how to critically evaluate their outputs. An AI recommendation should not be treated as a final verdict. The doctor must understand the tool's validation data, its known limitations, and its error rates before relying on its outputs in patient care.

Third, hospitals and health systems should maintain transparency about which AI tools they use and make this information available to patients on request. The ABDM (Ayushman Bharat Digital Mission) framework, which is building a national digital health ecosystem, could serve as a natural vehicle for standardising disclosure requirements across healthcare providers.

Fourth, AI developers operating in India should be required to publish evidence of clinical validation on Indian patient populations before their tools are deployed at scale. Self-reported accuracy metrics from foreign trials are not sufficient justification for broad use in a clinically and genetically diverse country like India.

The Ethical Dimension Beyond Law

The legal question of who is liable when an AI-influenced decision goes wrong is important, but it is not the only question that matters. There is a deeper ethical question about the kind of medicine India wants to build as it rapidly embraces digital health.

Medicine has always involved an inherent power imbalance between the patient and the provider. The patient is often unwell, anxious, and dependent on the physician's knowledge. Informed consent was specifically designed to counteract that imbalance by giving patients agency and voice in decisions about their own bodies. Introducing AI into clinical care without transparent disclosure risks deepening that imbalance by adding another layer of complexity that the patient cannot see, question, or challenge.

For Indian patients, many of whom already navigate considerable barriers to healthcare access including affordability, literacy, and geographic distance, the introduction of invisible algorithms into their care carries additional risks. Health literacy remains a significant challenge across much of India. Patients who do not understand that an algorithm is influencing their diagnosis cannot meaningfully question or refuse that influence. This is not a theoretical concern. It is a real and present gap in the ethics of digital health in India today.

Medicircle recognises that as healthcare communication evolves, so does the responsibility to bring important conversations like this one to the forefront, helping doctors, hospitals, patients, and healthcare institutions stay informed about the rapidly changing landscape of technology and ethics in medicine.

Frequently Asked Questions

Q1: Is it legally required for Indian doctors to inform patients when AI is used in their diagnosis or treatment?

Currently, there is no explicit legal mandate in India requiring disclosure of AI involvement in clinical decision-making. However, under existing informed consent principles and NMC guidelines, doctors are expected to share material information with patients, and AI use in care may qualify as such information. Clear regulatory standards are still being developed.

Q2: Who is legally responsible if an AI tool gives a wrong recommendation that leads to patient harm in India?

Legal responsibility typically falls primarily on the treating physician, who is expected to critically evaluate any AI recommendation before acting on it. However, the AI developer and the deploying hospital may also face liability depending on the circumstances, including whether the tool was adequately validated and whether the institution followed responsible deployment practices.

Q3: How can patients in India find out whether AI is being used in their medical care?

Patients have the right to ask their treating doctor or hospital whether AI-based tools are being used in their diagnosis or treatment. While disclosure is not yet universally mandated, patients can and should ask this question directly. Hospitals following responsible digital health practices should be able to provide a clear and honest answer.

Q4: Are AI tools used in Indian hospitals validated for Indian patients?

Not always. Many AI tools used in India were developed and validated primarily on patient populations from Western countries. Validation on Indian patient populations remains limited for many tools. This is a significant concern that regulators, hospitals, and AI developers need to address urgently.

Q5: What role does the ABDM play in AI governance and patient data in India?

The Ayushman Bharat Digital Mission (ABDM) is building the foundational infrastructure for India's digital health ecosystem, including health ID systems, health records, and data exchange frameworks. While it does not yet have specific AI governance mandates, it provides a potential platform for standardising how AI tools are disclosed, validated, and regulated in Indian healthcare settings.

Resources

1. Indian Council of Medical Research (ICMR): Biomedical research guidelines and ethical standards for AI in healthcare in India.
2. National Medical Commission (NMC): Professional conduct, etiquette, and ethics regulations governing medical practice and patient consent in India.
3. World Health Organization (WHO): Global guidance on ethics and governance of AI for health, including principles of transparency and accountability.
4. Ayushman Bharat Digital Mission (ABDM): India's national digital health ecosystem framework covering health data, digital infrastructure, and healthcare technology governance.
5. Ministry of Health and Family Welfare (MoHFW): Government policies, digital health initiatives, and regulatory updates relevant to AI adoption in Indian healthcare.

Interlinking Keywords

AI in Indian healthcare, patient informed consent India, digital health ethics India, AI medical decision support, ABDM digital health, AI bias Indian patients, healthcare AI accountability, NMC consent guidelines, AI radiology India, responsible digital health

Last reviewed by:

Dr. Manthan Tripathi, Medicircle Editorial and Medical Advisory Team on 5, October 2026.

Disclaimer

This article is intended for educational and informational purposes only. It does not constitute medical advice, legal advice, or a clinical recommendation of any kind. Readers should consult a qualified and registered medical professional for any health-related concerns or clinical decisions. The information presented reflects publicly available knowledge and does not substitute for personalised professional guidance. Medicircle does not endorse any specific AI tool, technology company, hospital, or clinical product.