Authored By: Winkie K Molefe
University of South Africa
ABSTARCT
The rise of Artificial Intelligence (AI) in South African healthcare presents unprecedented opportunities for enhanced diagnostics, improved efficiency, and more effective patient management. At the same time, AI introduces complex challenges concerning medical negligence, accountability, and the preservation of professional judgment. South Africa faces a growing medical negligence crisis, characterised by chronic understaffing, deteriorating public health facilities, and a dramatic rise in multimillion-rand malpractice payouts.
This article examines how AI is likely to operate within this fragile landscape, drawing on South African constitutional and statutory frameworks, common-law principles governing negligence, and a selection of high-profile medicolegal cases. It also integrates comparative insights from the European Union AI Act, the United States FDA framework for AI-based medical devices, and international ethical guidance.
Ultimately, the article argues that AI must function as an assistive tool rather than a replacement for clinical reasoning, and that a robust South African regulatory regime, coupled with strengthened accountability, improved patient protection, and alignment with international best practice, is essential to mitigate risks and maximise the benefits of medical AI.
INTRODUCTION
Medical negligence in South Africa occurs when healthcare providers fail to meet the expected standard of care, resulting in harm or injury to patients. It is defined as the failure of a healthcare professional, such as a doctor, nurse, or hospital, to provide the standard of care that a competent professional would offer under similar circumstances.1 Such failures can lead to significant physical, emotional, or financial harm, and in severe cases, death.
South African medical negligence claims are governed by the law of delict, which requires the plaintiff to establish: (i) a legal duty of care owed by the healthcare provider, (ii) a breach of that duty, (iii) causation linking the breach to harm, and (iv) damages suffered by the patient. 2 Common examples include misdiagnosis or delayed diagnosis, surgical errors, medication errors, and negligence during pregnancy or delivery.
Patients in South Africa are protected by a legal and constitutional framework that affirms their right to adequate healthcare and the ability to seek redress for harm caused by medical negligence. Understanding these legal foundations, case law precedents, and statutory requirements is crucial for both patients and practitioners.
South Africa’s health sector is marked by rising demands, deteriorating infrastructure, and a severe shortage of health care professionals. Public hospitals remain overwhelmed by patient numbers, understaffed wards, outdated equipment, and systemic administrative failures. These structural weaknesses have contributed directly to a sharp rise in medical negligence claims. Between 2011 and 2016, claims against provincial departments increased by roughly 20% per year, and since January 2015, the Gauteng Department of Health alone has reportedly paid over R1.7 billion in settlement payouts across 185 cases of medical negligence. 3
The situation continues to worsen medical negligence claims increasing by an average of 23% per year between 2014 and 2020, and more than R4.12 billion was paid out across South Africa between 2020 and 2023.4 These costs significantly disturb healthcare budgeting, reducing funds available for essential services, modern equipment, and staff training.
Catastrophic incidents, such as the Life Esidimeni tragedy, which resulted in the deaths of 93 mental health patients following negligent transfers, underscore the systemic challenges and the urgent need for enhanced oversight and accountability. Against this backdrop, the integration of Artificial Intelligence (AI) into healthcare has accelerated. AI tools are increasingly used in radiology, diagnostics, predictive analytics, pathology, hospital triage systems, and even surgical decision-support.
These systems promise to reduce human error, facilitate early detection, and relieve overburdened medical personnel. However, AI is not infallible. Errors arising from biased data, algorithmic malfunction, insufficient contextual reasoning, or over-reliance by practitioners can cause severe patient harm. This article examines how AI will interface with South Africa’s legal framework governing medical negligence, focusing on duty of care, professional competence, liability allocation, and patient rights.
It incorporates South African constitutional provisions, statutory duties, the Patient Rights Charter, common-law negligence principles, and leading case law. Comparative insights from the EU AI Act, US FDA regulatory approach, OECD AI Principles, and global medical ethics guidelines further illuminate the challenges and opportunities. Ultimately, the article argues for a multi-layered regulatory approach to ensure that AI complements, rather than displaces, human clinical judgment.
Research Methodology
This research adopts a doctrinal and comparative legal methodology.
Primary sources include:
- the Constitution of the Republic of South Africa, 1996
- statutory frameworks including the National Health Act, Health Professions Act, Medicines and Related Substances Act, Consumer Protection Act, and POPIA
- common-law negligence principles
- leading South African court decisions on medical malpractice
- The Patient Rights Charter
Secondary sources include journal articles, medico-legal reports, press investigations, Department of Health briefings, and practitioner commentary. International frameworks, most notably the EU AI Act (2023), the US FDA SaMD guidance, OECD AI Principles, and WHO guidelines on AI in health, provide a comparative perspective.
BODY
The Landscape of Medical Negligence in South Africa
Under South African law, healthcare practitioners owe a duty of care to their patients, derived from both common law principles of negligence and statutory requirements under the Health Professions Act 56 of 1974.5 The duty of care mandates that practitioners exercise reasonable skill, diligence, and competence in treating patients. Breach of this duty, resulting in harm, constitutes medical negligence, entitling patients to compensation. 6
The National Health Act 61 of 2003 further outlines institutional responsibilities, emphasising the provision of adequate healthcare services and the protection of patient rights.7 Hospitals and healthcare institutions are required to maintain standards that prevent foreseeable harm, including ensuring that staff are properly trained, facilities are adequately equipped, and clinical procedures adhere to established protocols.8
Rising Claims and Structural Failures, South Africa’s medicolegal environment is characterised by worsening outcomes. Medical negligence claims rose 23% annually from 2014–2020.9 As of March 2023, medicolegal claims against the State totalled R68 billion. 10 In 2022–2023 alone, R1.42 billion was paid out for claims involving brain-damaged infants.11 These figures reflect deeper structural issues,
including:
- chronic understaffing and brain drain of senior clinicians
- outdated equipment
- overcrowded hospitals
- administrative incompetence
- dysfunctional referral systems
- poor record-keeping
- lack of accountability
In many cases, negligent treatment leads to irreversible, catastrophic harm, especially in obstetrics, neurosurgery, and emergency medicine.
Hospital Horror Cases (DSC Attorneys and TGH Examples)
Birth-related neurological injuries constitute the largest category of medical negligence claims. In KwaZulu-Natal, a hospital’s failure to perform an emergency C-section, despite clear fetal distress, caused oxygen starvation and permanent impairment. Damages awarded: R12.9 million.12 Life Esidimeni Tragedy, between 2015 and 2016, 1,300 psychiatric patients were moved from Life Esidimeni to 27 unlicensed, ill-equipped facilities, resulting in at least 93 deaths from dehydration, malnutrition, hypothermia, and neglect.13 This remains one of the most severe systemic medical negligence events in democratic South Africa.
Surgical Negligence Wrong-site surgeries, retained surgical items, and incompetent post operative monitoring remain common sources of litigation. Private Sector Malpractice Negligence also occurs in private hospitals. A child who underwent repeated faulty neurosurgical operations in Mossel Bay suffered permanent cognitive and mobility impairments; damages reached R25 million.14
Individual Patient Narratives, The Think Global Health report describes Ms Bamby Manamela, whose uterus and kidney were damaged during negligent surgeries. Her medical and legal journey illustrates denial of responsibility by practitioners, delays in diagnosis, exploitation by legal representatives, and severe mental consequences. These narratives emphasise the human cost of system-wide failures. 15
Legal Framework Governing Medical Negligence in South Africa
Constitutional Rights
The Constitution establishes clear health-related entitlements, Section 27(1)(a): everyone has the right to access healthcare services.16 Section 12(2): security in and control over one’s body.17 Section 10: the right to human dignity.18 Section 195: Public administration must be efficient, accountable, and transparent.19 These constitutional rights underpin claims of medical negligence, ensuring that patients have both a right to quality care and legal avenues to claim compensation when these rights are violated.
Patient Rights Charter
The PRC affirms that patients have the right to dignity, privacy, and confidentiality. The right to be informed consent. The right to quality care and appropriate. The right to seek compensation for harm. The Charter is aligned with both the Constitution and statutory law, ensuring that healthcare providers are legally accountable for patient welfare and failure to uphold the Charter may strengthen negligence claims.20
Statutory Framework
(a) National Health Act 61 of 2003
The NHA establishes the duties and responsibilities of healthcare providers and the Department of Health. It emphasises access to healthcare services, quality assurance in healthcare delivery and monitoring and evaluation of health services.21
(b) Health Professions Act 56 of 1974
The HPA governs the conduct and professional standards of healthcare practitioners. It mandates registration, ethical conduct, and disciplinary procedures. Crucially, it reinforces the duty of care owed by health professionals and provides for oversight by the Health Professions Council of South Africa (HPCSA), which investigates complaints of negligence and misconduct.22
(c) Medicines and Related Substances Act 101 of 1965
This Act regulates medicines and medical devices, including AI systems classified as medical devices. It allows for the incorporation of AI to enhance the functionality of machines, monitor patient data, and improve diagnostics. Compliance with this Act ensures that AI-assisted devices meet safety and efficacy standards, supporting the standard of care required in healthcare.23
(d) Protection of Personal Information Act (POPIA)
POPIA protects the confidentiality of patient information, which is particularly relevant in AI applications that use patient data for training algorithms. Healthcare providers must ensure that AI systems process personal health information lawfully, maintain privacy, and secure consent from patients. 24
Common-Law Test for Negligence
South African courts apply a variation of the Bolam test: did the practitioner act in accordance with the standard expected of a reasonably competent practitioner in their field? Courts also consider, foreseeability of harm, reasonableness of precautions, causation and expert evidence. Crucially, AI reliance does not remove the practitioner’s duty to exercise independent clinical judgment. 25
Key Case Laws
(a) NVM obo VKM v Tembisa Hospital (2020)
Failure to monitor fetal distress caused severe brain injury. Emphasised institutional liability.26
(b) ZASCA 2022/186
Clarified shared liability where clinical oversight failures combine with system deficiencies.27
In SS v MEC for Health (Zithulele Hospital, 2014), the plaintiff, Ms S.S., claimed damages on behalf of her child, E.S., who suffered brain damage due to alleged negligence during labor at Zithulele Hospital. The plaintiff presented evidence of repeated failures in monitoring labor, delayed interventions, and improper management, culminating in hypoxic-ischemic injury.
The court found the defendant negligent. The hospital staff failed to act according to the guidelines, foresee potential harm, and implement protective measures. The claim succeeded on a balance of probabilities. This case establishes that delayed intervention and inadequate monitoring constitute actionable negligence, reinforcing the duty of care owed to both mothers and newborns.28
In AM v MEC for Health (Nelson Mandela Academic Hospital, 2017)
Plaintiff ALJ M., representing her minor child L…, claimed that negligent monitoring and delayed Caesarean section during labor led to L…’s cerebral palsy. The hospital conceded that the staff’s actions were substandard.
The court ruled in favour of the plaintiff, holding the MEC vicariously liable for the staff’s negligence and ordering payment of damages. The case underscores the vicarious liability of health departments and the importance of timely obstetric intervention.29
AI in Healthcare:
Opportunities and Risks
Opportunities AI can improve in South Africa, analyse imaging with high precision and detect early-stage diseases. Predict patient deterioration and optimise resource allocation, also reduce administrative burdens. This is especially valuable in South Africa’s overstretched public hospitals.
Risks
Algorithmic Error AI may misdiagnose due to biased or incomplete datasets, poor contextual understanding and error propagation. Over-reliance, systems may defer critical judgment to AI systems. Liability confusion: who is responsible when AI contributes to harm? The developer? The hospital? the state? The clinician?
Comparative International AI Regulation
United Kingdom (NHS Duty of Care) UK courts hold healthcare providers liable when their actions fall below expected professional standards.30 The Bolam test assesses whether the practitioner acted in accordance with a responsible body of medical opinion. AI-assisted devices must comply with MHRA medical device regulations, mirroring South Africa’s Medicines Act. 31
European Union (AI Act, Medical Devices) AI systems in healthcare are classified as medical devices. Providers must ensure safety, effectiveness, and transparency.32 Liability arises if AI contributes to patient harm due to inadequate design, monitoring, or misuse. These global models provide templates South Africa can adapt. 33
AI and Medical Negligence: Analytical Challenges in South Africa
Does AI Change the Standard of Care? Courts will likely find that AI can inform decisions and cannot replace clinical judgment. Also, reliance on AI must be reasonable. AI can assist in diagnostics, monitoring, and treatment planning. However, it must comply with the Medicines Act (as medical devices) and POPIA (data protection). Hospitals remain vicariously liable for AI errors if staff fail to oversee or implement AI recommendations appropriately. Patients’ rights, as outlined in the PRC, require informed consent and confidentiality in AI-assisted healthcare.
AI and the Duty to Keep Updated: AI introduces a duty of technological proficiency. Practitioners must understand AI limitations and question abnormal outputs, and document decision-making. Failure may constitute negligence. Record-Keeping and Explainability AI must be transparent, auditable and explainable. Opaque systems make litigation more complex.
Risk of Worsening Inequality: If AI is deployed unevenly across public/private sectors, it may deepen treatment disparities.
Recommendations and Way Forward
South African AI-Health Regulatory Framework
- Strengthen oversight: HPCSA and health departments should actively monitor AI integration.
- Standardise AI protocols: Guidelines for AI use in diagnostics, monitoring, and treatment.
- Education and training: Upskill healthcare staff in AI literacy.
- Legal reform: Introduce specialist medical courts or mediation channels for negligence claims.
- Data protection compliance: Ensure patient consent for AI-assisted treatment and secure handling of health data.
Conclusion
Artificial Intelligence holds enormous potential to support South African healthcare, especially in an environment characterised by understaffing, resource shortages, and rising medical negligence claims. However, without careful regulation, transparency, and robust oversight, AI may aggravate rather than alleviate systemic weaknesses. South Africa’s constitutional guarantees, statutory duties, the Patient Rights Charter, and common-law negligence principles already impose clear obligations on healthcare practitioners and institutions. As AI becomes more embedded in medical decision-making, these obligations will evolve but not disappear.
To harness AI effectively, South Africa must adopt a comprehensive legal and ethical framework aligned with international standards but responsive to local realities. AI must remain a tool under human supervision, not a replacement for the professional judgment that lies at the heart of safe, dignified, and lawful medical care. Only through responsible integration can AI contribute meaningfully to reducing medical negligence and strengthening the right to health for all.
OSCOLA BIBLIOGRAPHY
Legislation
- Constitution of the Republic of South Africa, 1996
- National Health Act 61 of 2003
- Health Professions Act 56 of 1974
- Medicines and Related Substances Act 101 of 1965
- Consumer Protection Act 68 of 2008
- Protection of Personal Information Act 4 of 2013
Cases
NVM obo VKM v Tembisa Hospital (2020)
Life Esidimeni Arbitration Award (2018)
ZASCA 2022/186
A.M v MEC for Health (401/2021) [2024] ZAECBHC 19 (6 August 2024).
S.S v MEC for Health (35/2019) [2023] ZAECMHC 8 (28 February 2023).
International Instruments
- European Union AI Act (2023)
- US FDA, Guidance on Software as a Medical Device
- OECD AI Principles
- WHO, Ethics and Governance of AI for Health (2021) Secondary Sources
- DSC Attorneys, Medical Malpractice Rift in South Africa: SA’s Hospital Horror Stories
- Think Global Health, “South Africa’s Medical Negligence Crisis”
- Larisse Prinsen, commentary in TGH interview
Books and Articles
Larisse Prinsen, ‘South Africa’s Medical Negligence Crisis’ Think Global Health https://www.thinkglobalhealth.org/article/south-africas-medical-negligence crisis accessed 8 November 2025
‘Medical malpractice resources: S.A. hospital horror stories’ D.S.C Law https://www.dsclaw.co.za/medical-malpractice-resources/s-a-s-hospital-horror stories/ accessed 8 November 2025
South African Health Reviews, ‘Health Legislation’https://www.hst.org.za/ publications/South%20African%20Health%20Reviews/07%20%20Chapter%201.%2 0Health%20Legislation.pdf accessed 8 November 2025
Foundations of Medical Law, University of Western Cape https:// law.uwc.ac.za/programme/foundations-of-medical-law/ accessed 8 November 2025
1 Burger Huyser Attorneys.
2 Legal Advice South Africa.
3 DSC Attorneys, SA’s Hospital Horror Stories Article (2016).
4 Think Global Health, “South Africa’s Medical Negligence Crisis” Article (2023).
5 Ibid.
6 Gary Austin quoted in Think Global Health (2023).
7 National Health Act 61 of 2003, ss 1.
8 Ibid.
9 DSC Attorneys (n 3).
10 DSC Attorneys (n 3).
11 DSC Attorneys (n 3).
12 DSC Attorneys (n 3).
13 DSC Attorneys (n 3).
14 DSC Attorneys (n 3).
15 Think Global Health (n4).
16 Constitution of the Republic of South Africa, 1996, 27 (1) (a).
17 Constitution of the Republic of South Africa, s 12 (2).
18 Constitution of the Republic of South Africa, s 10.
19 Constitution of the Republic of South Africa, 195.
20 Department of Health, Patient’s Rights Charter (Republic of South Africa.
21 National Health Act, s2.
22 Health Professions Act 56 of 1974, 3.
23 Medicines and Related Substances Act 101 of 1965.
24 Protection of Personal Information Act 4 of 2013.
25 Bolam v Friern Hospital Management Committee [1957] WRL 582 (CA).
26 NVM obo VKM v Tembisa Hospital and Another CCT202/20.
27 ZASCA 2022/186.
28 S.S v MEC for Health (35/2019) [2023] ZAECMHC 8 (28 February 2023).
29 A.M v MEC for Health (401/2021) [2024] ZAECBHC 19 (6 August 2024).
30 EU AI ACT (2023).
31 OECD AI Principles (2019).
32 US FDA, Guidance on Software as a Medical Device (2022).
33 WHO, Ethics and Governance of AI for Health (2021).
