Artificial Intelligence In Healthcare: Who Is Liable For Medical Negligence?

Author(s): Dr. S. Krishnan and Kusum Kumari

Paper Details: Volume 4, Issue 5

Citation: IJLSSS 4(5) 18

Page No: 224 – 244

ABSTRACT

Artificial Intelligence has emerged as one of the most revolutionary technologies in modern healthcare. From detecting diseases at an early stage to assisting in complex surgical procedures, AI offers numerous benefits. However, reliance on AI systems creates legal and ethical concerns, particularly when medical errors occur.

Traditional principles of medical negligence are based on human actions and professional standards of care. The increasing use of AI challenges these principles because machines now play a significant role in medical decision-making. This article discusses the legal framework governing medical negligence, examines the potential liability of doctors, hospitals, and AI developers, and analyses judicial precedents relevant to medical negligence. It further explores the need for a robust regulatory framework to address liability issues arising from AI-assisted healthcare.

Keywords: AI, Healthcare, Liability, Medical Negligence, Ethics, Medical Oath

INTRODUCTION

Robots armed with artificial intelligence (AI) will replace doctors by 2035, according to at least one “legendary Silicon Valley investor,” and in some cases, AI is already better than human doctors. Today, for example, AI can (1) “look at brain scans of people who are exhibiting memory loss and tell who will go on to develop full-blown Alzheimer’s disease and who won’t,” (2) allow hospitals “to predict the likelihood of a cardiac arrest in 70 percent of occasions, five minutes before the event occurs,” and (3) save lives and speed hospital discharge by improving treatment for “a deadly blood infection called sepsis.”

AI is being incorporated into health care worldwide. By 2030, researchers predict that AI may affect up to 14% of global domestic product with half of this effect coming from improvements in productivity. AI will transform healthcare by “deriving new and important insights from the vast amount of data generated during the delivery of health care every day.” AI can quickly and cost-effectively analyze previously unscalable data sets (like electronic health record data, medical images, laboratory results, prescriptions, and demographics) “to make predictions and recommend interventions” in patient care. However, “AI is only as good as the humans programming it and the system in which it operates.” Generally speaking, AI is defined as computer technology designed to perform tasks like, or better than, humans. AI mimics human intelligence using computer algorithms that learn from existing data by incorporating statistics and mathematics on a larger scale than generally possible for humans. An algorithm is simply sets of computer software code with instructions for the computer to perform certain tasks like recognizing patterns, reaching a conclusion, or predicting future behavior.

The growing prevalence of AI in the healthcare industry is fundamentally changing the dynamics of medical decision-making processes, especially in the realms of diagnostic tools and robot-assisted surgeries. Although these technological innovations provide significant improvements in terms of accuracy and efficiency, they undermine the foundations of negligence law, which relies heavily on human error and accountability.

Artificial Intelligence (AI) is transforming the healthcare sector by assisting doctors in diagnosis, treatment planning, robotic surgeries, drug discovery, and patient monitoring. While AI has significantly improved efficiency and accuracy in healthcare, it has also raised a crucial legal question: Who is liable when AI causes harm to a patient?

Under medical negligence law, there must exist a direct doctor-patient relationship in order for the physician’s liability to be established. In India, such liabilities have been determined based on standards of reasonable care, as outlined in the case of Dr. Laxman Balkrishna Joshi v. Trimbak Bapu Godbole. According to this ruling, the standard of medical practice requires that doctors maintain a high degree of skill and care within their professional duties. Nevertheless, in cases involving AI in the medical field, fault and causation cannot easily be established, since the process of decision-making becomes more distributed.

The issue becomes complex because AI systems are designed, developed, and used by multiple stakeholders, including software developers, hospitals, healthcare institutions, and medical professionals. If an AI-based diagnostic tool misdiagnoses a disease or a surgical robot malfunction during an operation, determining liability for medical negligence becomes challenging.

This article examines the concept of medical negligence in the era of artificial intelligence, analyses relevant legal principles, and explores the possible liability of various stakeholders.

Understanding Medical Negligence

Medical negligence refers to a breach of duty of care by a healthcare professional that results in injury or harm to a patient.

The essential elements of medical negligence are:

1. Existence of a duty of care.
2. Breach of that duty.
3. Causation between the breach and injury.
4. Actual damage suffered by the patient.

In healthcare, doctors are expected to exercise reasonable skill, care, and competence while treating patients. Failure to meet the accepted standard of medical practice may result in liability.

AI IN HEALTHCARE: A NEW PARADIGM

Artificial Intelligence refers to computer systems capable of performing tasks that normally require human intelligence. In healthcare, AI is used for:

• Disease diagnosis and prediction.
• Medical imaging analysis.
• Robotic-assisted surgeries.
• Personalized treatment planning.
• Drug development and research.
• Virtual health assistants.
• Remote patient monitoring.

Examples include IBM Watson Health, AI-based radiology tools, and robotic surgical systems such as the Da Vinci Surgical System.

Although these technologies improve healthcare outcomes, they are not immune from errors. AI systems may produce inaccurate results because of defective algorithms, biased data, software bugs, or improper use by medical professionals.

ARTIFICIAL INTELLIGENCE AND ROBOTICS IN HEALTHCARE: A CONCEPTUAL UNDERSTANDING

Artificial intelligence (AI) in healthcare can be defined as any computer technology that is designed for carrying out cognitive tasks usually done by humans, such as diagnostics, treatment, prediction and decision-making through machine learning, robots, and algorithms.9 This definition encompasses preventive care, diagnostics, surgeries, and administration in hospitals among other areas of health care practice.

DIAGNOSTIC AI

Diagnosis using artificial intelligence is becoming common practice within the realms of radiology, pathology, ophthalmology, and oncology. They use machine learning to diagnose diseases through image recognition and patient information. This increases efficiency within diagnosis and may even improve disease detection.

PREDICTIVE ANALYTICS

Predictive analytics uses machine learning algorithms to predict outcomes such as disease progression, patient worsening, treatment success, and public health trends. The downside to these algorithms is that they can have a tendency to embed algorithmic bias into their processes from bad data input.

ROBOTIC SURGERY

Robotic surgery helps minimize surgeries by allowing for less invasive surgeries to be done with increased accuracy, dexterity, and better recovery outcomes. However, technological malfunctions or errors may raise the complexity of any negligence lawsuit involving the robotic device.

CLINICAL DECISION-SUPPORT SYSTEMS (CDSS)

CDSS helps guide physicians through treatment decisions with advice and medication alerts. This may become problematic when physicians rely heavily on these systems due to their nature of increasing physician liability.

BENEFITS AND DRAWBACKS

The implementation of artificial intelligence and robotics in healthcare presents various benefits such as improved diagnostic capabilities, increased efficiency, personalized treatment, minimization of mistakes, and improved access to healthcare. Such technological innovations may be especially advantageous for resource-limited healthcare settings due to their fast and consistent performance. However, the deployment of AI is associated with significant legal and ethical problems, such as algorithmic bias, invasion of privacy, cyber security issues, software malfunctions, and difficulty in identifying the responsible party when something goes wrong. Responsibility distribution among several parties weakens negligence theory principles of human culpability.

BLACK-BOX ALGORITHMS AND LACK OF EXPLAINABILITY

One of the major issues in AI healthcare revolves around the “black-box” nature of the technology wherein no clear explanation accompanies the results. This lack of explainability prevents patients from giving valid consent, deprives patients of autonomy and the right to know about the recommendations, and reduces physicians’ liability for any AI recommendation or risk involved. For liability purposes, explainability is an important element in assessing whether there is breach of duty, whether the breach is foreseeable and what the cause of the injury is under negligence law. In other words, the absence of explainable AI means that current medico-legal frameworks are insufficient to govern AI liability.

USE OF LEGAL JARGON

Several legal concepts become relevant when discussing AI-related medical negligence:

DUTY OF CARE

Duty of care exists when there exists a doctor-patient relationship whereby it is incumbent upon any medical practitioner to act reasonably and with skill and diligence. Under Indian medical jurisprudence, it can be argued that medical professionals owe a duty of care in their decision to accept a particular case for practice, determine treatment, and administer the treatment with caution. The Supreme Court ruling in Dr. Laxman Balkrishna Joshi vs. Trimbak Bapu Godbole stated that for a medical practitioner to be regarded as competent, he must have a sufficient degree of skill and knowledge compared to others in the profession. However, this should not be taken as having the greatest skill or knowledge level.

BREACH OF DUTY

Breach is established where the defendant fails to adhere to the medical standard of care of an ordinarily skilled healthcare practitioner in similar circumstances. The Indian legal system employs the Bolam principle to assess breaches, analyzing whether the defendant’s actions conform to accepted practice among a responsible body of medical experts. This approach values the medical consensus more than the patient-oriented perspective, usually providing immunity to practitioners whose actions are supported by recognized medical opinions.

CAUSATION

Causation mandates that the breach be directly responsible for any injury sustained by the plaintiff. Generally, the “but for” rule is employed to prove whether any damage would have occurred had there been no negligence. Proving causation may be challenging within Indian medical malpractice laws owing to several contributing elements, prior illnesses, or ambiguous treatment procedures. The medical negligence laws in India are typically reliant on the conduct of physicians to establish causality.

VICARIOUS LIABILITY

• A legal doctrine under which an employer may be held liable for the wrongful acts of its employees performed during employment.

PRODUCT LIABILITY

• Liability arising from defects in products, including software and AI systems, that cause harm to users.

STANDARD OF CARE

• The degree of care and skill expected from a reasonably competent professional under similar circumstances.

FORESEEABILITY

• The ability to predict that certain conduct may result in harm.

WHO CAN BE HELD LIABLE?

1. LIABILITY OF DOCTORS

Doctors remain the primary decision-makers in most healthcare settings. If a physician blindly relies on an AI-generated recommendation without exercising independent medical judgment, the doctor may be held liable.

For example, if an AI diagnostic system incorrectly identifies a malignant tumour as benign and the doctor accepts the recommendation without further examination, resulting in harm to the patient, negligence may be attributed to the doctor.

The law generally expects doctors to use AI as an assistive tool rather than a substitute for professional judgment.

2. LIABILITY OF HOSPITALS

Hospitals may be held liable under the principle of vicarious liability if their employees negligently use AI systems.

A hospital may also be directly liable if it:

• Fails to properly maintain AI equipment.
• Uses untested or defective AI software.
• Neglects staff training regarding AI technologies.

Healthcare institutions have a responsibility to ensure that technological systems meet safety standards and are appropriately monitored.

3. LIABILITY OF AI DEVELOPERS AND MANUFACTURERS

Software companies and AI developers may be liable under product liability principles if harm results from:

• Defective algorithms.
• Programming errors.
• Insufficient testing.
• Failure to provide adequate warnings.

If an AI system consistently produces inaccurate diagnostic results due to a design defect, the manufacturer may be responsible for damages.The challenge lies in proving that the defect in the AI system directly caused the injury.

4. SHARED LIABILITY

In many situations, liability may be shared among multiple parties.

For example:

• -The AI developer creates a flawed algorithm.
• The hospital fails to test the system adequately.
• The doctor relies excessively on the AI output.

In such circumstances, courts may apportion liability among all responsible parties.

• The Proof: To establish medical negligence involving AI, the claimant must prove:

1. A duty of care existed.
2. The AI system contributed to the medical decision.
3. There was a breach of the required standard of care.
4. The breach caused injury.
5. Actual damages resulted.
6. Evidence may include:
7. Medical records.
8. Expert testimony.
9. AI system logs.
10. Software performance reports.
11. Internal hospital protocols.

Because AI systems often operate as “black boxes,” understanding how a particular decision was made can be difficult. This creates evidentiary challenges in litigation.

Difference Between Negligence and Strict Liability

Suppose a physician uses a medical AI in the treatment of a Black patient with cancer. The AI recommends an incorrect nonstandard drug dosage that the physician follows, and the patient’s condition worsens. As seen above, the physician may likely be held liable for causing injury to the patient. However, it turns out that the reason for the faulty AI recommendation was that the model was mainly trained on data from White patients. While the physician in this hypothetical scenario will likely incur liability for a bad patient outcome under current law, one key question still remains to be answered: Could the developer of the medical AI be likely held liable for negligence because the model was predominantly trained on data from White patients?

To establish a prima facie case for negligence, the plaintiff (here, the Black patient) must prove – by a preponderance of the evidence (that is, more than 50 per cent) – four elements: duty, breach, causation and damage. A successful negligence claim thus requires that the defendant (here, the AI developer) owes a legal duty to the patient and that this duty was accidentally breached, which caused the patient’s injury.

Up to the 1910s, injured consumers of flawed products were often unable to successfully sue manufacturers for negligence because they could not establish a duty of care due to a lack of contractual privity. Nowadays, courts no longer require privity for the existence of such a duty, and assume it. However, to recover for negligence, consumers still need to establish a breach of this duty, injury to them and actual and proximate causation between the breach and the injury.

Product liability – which is generally considered a strict liability of manufacturers for product defects – has evolved over time, and courts have established three types of product defects, namely (1) design defects, (2) manufacturing defects and (3) marketing defects.45 While a design defect is inherent and already exists before manufacturing the product, a manufacturing defect is a physical departure from the intended product’s design and occurs during its production or construction. Marketing defects refer to inadequate instructions or failures to warn consumers about possible risks associated with the use of the product. For instance, in our hypothetical example, a claim for a marketing defect may be given if the labelling of the AI did not include a warning that the model may likely not give reliable or correct recommendations when used in non-White patients. Obviously, such a model that has not been trained on a diverse patient population should not be placed on the market in the first place and may thus also trigger a design defect suit. When considering healthcare software, however, most courts have so far been hesitant to hold developers liable under product liability theories. The reason for this seems to be the assumption that such software is a clinical decision support tool that only gives recommendations and that it is the physician who ultimately decides. In other words, software has been interpreted as a service rather than a product. Thus, under current case law, it is likely that injured patients will have a hard time successfully suing developers of medical AIs under product liability. But a court’s shift to product liability would not be inconceivable in the future, considering that high-performing deep learning networks are increasingly being deployed in medicine, which are impossible or difficult for humans to understand (so-called black boxes).

An important distinction here is between a medical AI system that received marketing authorisation from the FDA and a medical AI system that is marketed without the need for FDA review. This distinction may be relevant for future court decisions regarding whether product liability applies in cases of healthcare software. The FDA does not regulate the practice of medicine (that is, services), but it does regulate medical devices, and if healthcare software is classified as such in a particular case, product liability is not outside the realm of possibility in the future. The distinction also matters because regulatory actions by the FDA may preempt state law, insulating some AI manufacturers from state-law tort claims.

CASE LAWS

1.DR. LAXMAN BALKRISHNA JOSHI V. TRIMBAK BAPU GODBOLE (1968)

The Supreme Court in the case of Dr Laxman Balkrishna Joshi Vs. Dr Trimbak Bapu Godbole and Anr. (1968) held that a doctor has a duty of care in deciding whether to undertake a particular case, a duty of care in deciding what treatment to give and a duty of care in the administration of that treatment. A breach of any of those duties gives a right of action for negligence to the patient. A doctor is expected to exercise a reasonable degree of care, neither the very highest nor a very low degree of care and competence judged in the light of the particular circumstances of each case is what the law requires.

The significant ruling laid down the principles of doctrine in relation to the duty owed by the doctors to their patients in India, stating that there is an obligation to exercise a reasonable degree of professional care in rendering treatment to their patients. Negligence cannot be established if it is due to error of judgment; it is only when the medical personnel fail to reach the standard set by other competent persons that the said act becomes negligence.

2. INDIAN MEDICAL ASSOCIATION V. V.P. SHANTHA (1995)

The Supreme Court held that medical services fall within the ambit of consumer services under the Consumer Protection Act. Patients can seek compensation for negligence committed by healthcare providers.

Significance: This case strengthened patients’ rights and established accountability in medical services.

3. JACOB MATHEW V. STATE OF PUNJAB (2005)

The Supreme Court clarified that a medical professional is liable only when negligence is gross and falls below the standard expected from a competent practitioner.

Significance: The case established the standard for determining medical negligence in India.

4. BOLAM V. FRIERN HOSPITAL MANAGEMENT COMMITTEE (1957)

The court held that a doctor is not negligent if their actions conform to a practice accepted as proper by a responsible body of medical professionals.

Significance: The famous “Bolam Test” remains influential in assessing professional negligence.

CONSUMER PROTECTION ACT, 2019

The Consumer Protection Act, 201927 provides remedy to patients by identifying the provision of medical care as consumer goods wherein failure in care can provide grounds for compensation. In light of the case of Indian Medical Association v. V.P. Shantha, patients can file complaints against their negligent caregivers to consumer tribunals. It has product Liability provisions, availability of remedy for deficiencies in medical services, allowing access to the dispute resolution and greater access for consumers. But this legislation focuses mainly on conventional medicine and does not cater to injuries inflicted through AI-assisted medical procedures.

LIMITATIONS OF TRADITIONAL DOCTRINE

Though the medical negligence doctrine is strong in theory within conventional settings, there are numerous shortcomings in India’s traditional medical negligence framework:

1. Focus on Human Agency – Traditional frameworks presume direct involvement of the physician in decision-making processes, hence rendering them unsuitable for application in AI technology-assisted healthcare.

2. Insufficient Incorporation of Product Liability – Defective algorithms, faulty robots, and software issues associated with medical treatment are not sufficiently covered by the current laws.

3. Insufficient Attention to Patients’ Right to Self-Determination – The Bolam test may undermine patients’ autonomy due to professional consensus-based approach.

4. Challenges of Proving Causation – Establishing the link between medical decisions and patient outcomes becomes more complicated with increasingly complex technological infrastructure.

5. No Specific Regulation – There is no comprehensive legislation governing AI healthcare liability within India.

Traditionally, the Indian medical negligence jurisprudence has offered physicians accountability through concepts of duty, breach, causation, and compensatory damages. The extensive reliance on physician-related standards such as Bolam test and Bolitho rule, coupled with insufficient modernization, makes the traditional legal framework obsolete in the rapidly evolving environment. With the emergence of new healthcare technologies, it is necessary to reformulate legal doctrines in order to protect patient autonomy while maintaining accountability.

LEGAL ISSUES CREATED BY ARTIFICIAL INTELLIGENCE IN HEALTHCARE

The application of artificial intelligence technology disrupts the traditional framework for medical malpractice litigation. The existing system of medical negligence revolves around the concept of human decision-making and the associated elements of duty, breach, and foreseeability. Yet, the adoption of new technologies creates a situation in which decisions can be made by an artificial agent without a traceable involvement of any physician. Thus, the existing legal structure is unable to allocate blame in situations involving algorithmic errors. Consequently, there is a need to review the existing approach to determining liability.

DIFFUSION OF LIABILITY

One of the primary legal issues created by the introduction of AI into healthcare practice is its diffusive nature. In cases where patients suffer harm caused by AI-driven diagnostic tools, surgical robots, or predictive systems, liability could be allocated to physicians, hospitals, software developers, manufacturers, providers of data and Governmental agencies In such cases, it would be difficult to determine who should be held responsible for patient injuries since not all actors possess control over medical outcomes.

INDEPENDENT ERROR AND ALGORITHMIC DECISION

Making AI technologies may create an independent error in diagnosing illness, suggesting treatment options, or performing procedures without continuous supervision from the doctor. Autonomous or semi-autonomous technology is incompatible with negligence laws since the error may arise without any negligence on the part of the human decision-maker.

Illustrative Examples Include:

A. Misdiagnosis through machine learning algorithms

B. Failure of surgery due to malfunctioning robots

C. Improper prediction outcomes due to faulty algorithmic operations

Here, deciding the liable party among physicians, programmers, manufacturers, or hospitals is legally ambiguous. The existing tort regime fails to consider liability for autonomous technological errors.

MANUFACTURER LIABILITY AND PRODUCT LIABILITY

Law As AI technologies become more advanced, medical instruments are increasingly regarded as products rather than services. In such scenarios, manufacturers will have to contend with questions about their liability for injuries caused by their products, including software and hardware defects and design flaws. The following would be the Legal Issues:

I. Whether an AI software considered a defective product?

II. Should strict liability be applied to autonomous medical technology?

 III. How do software updates impact liability?

In India, there is inadequate legal guidance within the scope of the Consumer Protection Act, 2019 regarding product liability in AI technology.

INSTITUTIONAL LIABILITY IN HOSPITALS

The facilities in hospitals are an important part in introducing, monitoring, and integrating AI applications into medical practice. In this regard, hospitals can be held liable when they fail to introduce safe AI application, train properly medical personnel, maintain AI tools and their software, consider the risks associated with technology and protect patient data The risk of hospital liabilities due to negligent use of technology and medical devices becomes ever more serious. The introduction of AI technology makes the responsibility extend from individual medical negligence to corporate liability.

SOFTWARE DEFICIENCIES AND TECHNICAL FAILURES

One of the hardest aspects of liability for AI technology is that of software deficiency. It can be caused by, coding errors, algorithm mistakes, improper testing, poor system integration, training data impropriety and failure to update. Different from medical malpractice, software deficiencies do not depend on physician conduct at all, so special considerations are needed to resolve them properly. There are no specific rules regulating AI technologies, which makes litigation even harder.

DATA BIAS AND DISCRIMINATORY OUTCOMES OF AI

The Artificial intelligence algorithms are heavily reliant upon data used during the training process and optimization. A biased dataset might lead to discriminatory results, causing additional harm to vulnerable populations due to race, gender, economic condition, or geographical discrepancy. The following would be the Legal Issues:

1. Discriminatory diagnosis
2. Unequal distribution of treatment
3. Misrepresented minorities
4. Structural injustice reinforcement

The data bias creates issues of constitutional, ethical, and consumer rights law, especially when applied to equality-based legislation and medical patient rights.

CYBERSECURITY THREATS AND DATA PRIVACY

The Healthcare AI systems handle extremely sensitive information about the patient’s health, which puts them at risk of cyber attacks, hacking, ransomware, and data corruption. Cybersecurity threats put at risk on patient privacy, accuracy of treatment, operation of the system and trust of institutions There is potential for legal liability, if a cyber breach leaves the patient susceptible to physical, mental, and financial danger. With an increasing presence of integrated healthcare systems, cybersecurity becomes the core principle of AI liability regulation.

EXPLANATORY GAP, BLACK BOXES, AND REGULATORY FAILURES

One of the key problems in the realm of artificial intelligence in the medical field is that of black boxes, which refers to medical algorithms whose output lacks any degree of explanatory power as to the manner in which such a conclusion was arrived at. As opposed to clinical diagnosis and other decision making processes carried out by doctors, there is no justification for the decisions taken through such processes. Such an explanatory gap poses substantial problems in the area of medical negligence due to the difficulty of assessing negligence, standards of care, foreseeability, and breach in light of the inability of these processes to be critically analyzed by the courts. Such opacity may also undermine the principle of informed consent whereby doctors lack any reasonable way of explaining the risks associated with their decisions made through AI technology. In India, the problem is further exacerbated by structural failures.

For example:

1. There is no special healthcare laws for AI in India
2. The presence of insufficient modification of product liability for AI applications in medicine
3. The continued persistence of human-centered negligence criteria
4. An insufficient regulatory supervision
5. The nonexistence of obligatory algorithmic audit and there is no enforceable requirements for transparency

While the Consumer Protection Act, 2019 provides partial solutions, it falls short in addressing damages caused by autonomous systems, software errors, and algorithmic bias. Consequently, there is a regulatory gap that leaves patients unprotected and healthcare providers liable under unclear circumstances. Thus, it becomes necessary to ensure explainability and regulatory reforms to create an accountable AI governance framework. India needs to adopt a hybrid regulatory model incorporating medical negligence, product liability, institutional accountability, and AI regulations.

COMPARATIVE JURISPRUDENCE: GLOBAL PERSPECTIVES ON LIABILITY FOR AI IN MEDICINE

UNITED KINGDOM: AUTONOMY OF THE PATIENT AND TRANSCENDING PROFESSIONAL PATERNALISM

In the United Kingdom, the medical jurisprudence landscape has advanced beyond physician centered negligence principles, embracing patient autonomy more effectively. The Montgomery v. Lanarkshire Health Board decision by the UK Supreme Court ruled against undue dependence on the classic Bolam test in cases involving information sharing and consent. The court ruled that physicians have a duty to guarantee patients are fully informed of all relevant risks and possible alternatives to their treatment. In the same case, there was a shift in medical negligence jurisprudence toward the principle of patient autonomy. This decision is highly relevant in AI-based medicine because the black box nature of the algorithm may prevent physicians from adequately informing their patients of the associated risks and limitations. Therefore, AI-based systems may constitute a violation of the patient autonomy principle.

EUROPEAN UNION: EU AI ACT AND RISK-BASED REGULATION

The European Union stands out in the governance of AI through its proposed Artificial Intelligence Act, which forms part of one of the world’s first comprehensive regulatory laws on artificial intelligence.¹ Rather than relying on the traditional approach of addressing liability only once harm is done, the European Union employs a precautionary risk-based regulation that aims to regulate AI technology according to the levels of risk posed to patients’ safety and health, as well as their fundamental rights. In this regard, AI technologies can be categorized as follows:

1. Unacceptable-risk technologies – In the EU framework there is a complete ban owing to the significant risks posed
2. High-risk technologies – There is a stringent regulations are required for the cases on high risked technological applications
3. Limited-risk technologies – It is very important to make the limited risk technology as mainly transparent obligations apply
4. Minimal-risk technologies – In this case, lightest regulation applies.

In the field of Medical AI, especially those employed for diagnostic purposes, robotic surgeries, treatment planning, and decision support, falls under the category of high-risk AI technologies owing to their direct impact on people’s health and medical procedures.

REGULATORY REQUIREMENTS OF HIGH-RISK HEALTHCARE AI TECHNOLOGIES BY THE EU

Some of the regulatory requirements for high-risk medical technologies, under the EU AI Act, include the following:

1. There is a requirement for conformity assessments before marketing the technology
2. The mandatory disclosure requirements where high risk medical technology is being used.
3. There is also a human supervision requirement to keep an eye and to maintain the accountability.
4. There is clear algorithmic and technical documentation when AI is being used in high risked medical procedures for the audits and records maintenance.
5. There is post-market monitoring, risk management and data governance to ensure that AI is being used with maintainability and a human supervision.

This approach not only allows for a pro-active approach that enables re-implementation monitoring of AI-based health care applications but also post-implementation monitoring of these technologies. This framework focuses on issues such as safety, liability, and accountability rather than taking a re-active approach to the issue of litigation.

HOW CAN INDIA ADOPT THESE STRUCTURES?

The frameworks set out by the UK and EU in addressing the use of artificial intelligence in medicine can offer considerable assistance to India as it begins to grapple with the challenges posed by the increasing presence of artificial intelligence in healthcare. The current medico-legal framework in India, which is still heavily based on negligence law centered on physicians, must adapt to the novel technology-related challenges brought about by AI-based medical technologies.

1. To begin with, India needs to establish robust patient autonomy guidelines akin to those provided by the UK in Montgomery v. Lanarkshire Health Board. Here, physicians must guarantee that their patients have been fully informed not just about general medical risks, but also risks related to AI-based medical technologies. By establishing stringent informed consent rules, patients’ rights in a technology-dependent healthcare environment will be preserved.
2. Secondly, India needs to establish informed consent guidelines concerning the use of artificial intelligence in medicine. Patients must be informed whether their diagnoses and treatment plans have been influenced by artificial intelligence technologies in any way.
3. Thirdly, employing a risk-based approach similar to the European Union’s AI Act would permit India to categorize medical AI applications based on how much they affect patient safety. Those that fall under high risk, such as diagnostic algorithms, robots in surgery, and clinical decision support systems, should be more rigorously assessed legally compared to low-risk applications.
4. Fourthly, India should also implement mandatory pre-market approval procedures for high-risk medical AI technology to guarantee that all high-risk technologies have undergone appropriate assessment concerning safety, efficiency, and transparency prior to their deployment.
5. Fifthly, post-market supervision must also take place since AI applications can keep learning and adapting even after they have been introduced into practice. Consequently, regulatory bodies should supervise all systems’ operations for any signs of malfunctions, cybersecurity threats, drift in the learning process, and newly emerged bias issues.
6. Ultimately, it is necessary to introduce mandatory provisions for transparency and explainability in healthcare AI technology in India. It should be ensured that doctors, patients, and judges have a sufficient understanding of the principles on which AI makes medical decisions. Explainability is important not only from the perspective of ethics but also for the efficient functioning of the judicial process in negligence claims.

In light of the above comparative analysis, it becomes clear that India could shift away from the obsolete negligence model centered around doctors’ responsibility and establish an up-to date regulatory approach involving patient autonomy, prevention, technical accountability, and institutional controls.

CHALLENGES IN DETERMINING LIABILITY

• Lack of Transparency: Many AI systems operate through complex algorithms that are difficult to interpret.
• Absence of Specific Legislation: India currently lacks a comprehensive legal framework specifically regulating AI liability in healthcare.
• Multiple Stakeholders: Determining responsibility becomes difficult when several actors contribute to the final outcome.
• Evolving Standards: The standard of care for AI-assisted medicine is still developing, making judicial assessment challenging.
• The Need for Legal Reform: As AI becomes increasingly integrated into healthcare, governments must establish clear legal guidelines.

Necessary reforms include:

• Dedicated AI healthcare regulations.
• Mandatory safety and testing standards.
• Transparency requirements for AI algorithms.
• Allocation of liability among stakeholders.
• Ethical oversight mechanisms.
• Data protection and patient privacy safeguards.

A balanced regulatory framework can encourage innovation while protecting patient rights.

CONCLUSION

Artificial Intelligence has the potential to revolutionize healthcare by improving diagnostic accuracy, reducing costs, and enhancing patient outcomes. However, AI also creates significant legal challenges regarding accountability and medical negligence.

Under existing legal principles, doctors, hospitals, and AI developers may all bear responsibility depending on the circumstances of the case. Courts are likely to assess whether each stakeholder fulfilled their respective duties and whether their actions contributed to the patient’s injury.

The future of healthcare will increasingly involve collaboration between humans and machines. Therefore, legal systems must evolve to ensure that technological advancement does not come at the cost of patient safety and justice. A clear and comprehensive regulatory framework is essential for balancing innovation with accountability in the age of artificial intelligence.

REFERENCES

1. World Health Organization, Ethics and Governance of Artificial Intelligence for Health 15-18 (2021).
2. Hisham Jadallah Mansour Shakhatreh & Farouq Ahmad Faleh Alazzam, Legal Frameworks for AI-Driven Diagnostics and Robotic Surgery: Addressing Medical Malpractice and Accountability, 33 Acta Informatica Medica 248, 249-50 (2025)
3. Francesco De Micco et al., Robotics and AI into Healthcare from the Perspective of the European Union Regulatory Framework: How to Fill the Gap Between Technology and Existing Liability Systems?, 11 Frontiers Med. 1428504, 5-6 (2024)
4. Aniek F. Markus et al., The Role of Explainability in Creating Trustworthy Artificial Intelligence for Health Care: A Comprehensive Survey of the Terminology, Design Choices, and Evaluation Strategies, 3 NPJ Digit. Med. 161, 3-4 (2021)
5. Consumer Protection Act, No. 35 of 2019.
6. Matúš Mesarčík et al., Stance on the Proposal for a Regulation Laying down Harmonised Rules on Artificial Intelligence – Artificial Intelligence Act, (2022), https://doi.org/10.31235/osf.io/yzfg8.
7. Montgomery v. Lanarkshire Health Bd., [2015] UKSC 11, [2015] A.C. 1430 (U.K.)
8. Bharatiya Nyaya Sanhita, 2023, No. 45 of 2023 (India).
9. Dr. Laxman Balkrishna Joshi v. Trimbak Bapu Godbole, (1969) 1 S.C.R. 206 (India)
10. Jacob Mathew v. State of Punjab, (2005) 6 S.C.C. 1 (India)
11. Bob Kocher & Zeke Emanuel, Will Robots Replace Drs.?, THE BROOKINGS INST. (Mar. 5, 2019), https://www.brookings.edu/blog/usc-brookingsschaeffer-on-health-policy/2019/03/05/will-robots-replace-doctors
12. Daisy Yuhas, Doctors have trouble diagnosing Alzheimer’s. AI doesn’t, NBC NEWS (Oct. 30, 2017), https://www.nbcnews.com/mach/science/ doctors-have-trouble-diagnosing-alzheimer-s-ai-doesn-t-ncna815561 [https://perma.cc/6DJU-8S4B].
13. David Shimabukuro, et al., Effect of machine learning-based severe sepsis prediction algorithm on patient survival and hospital length of stay: A randomised clinical trial, 4 BMJ OPEN RESP. RES. 1 (2017), https://bmjopenrespres.bmj.com/content/4/1/e000234 [https://perma.cc/R54D-8QV4].
14. W Nicholson Price II, ‘Artificial Intelligence in Health Care: Applications and Legal Implications’ (2017) 14 The SciTech Lawyer 10, 11.