New Technologies in International Law / Tymofeyeva, Crhák et al.

Developing nations may also solve the issue of liability by dealing with the black-box model problem by mandating increased transparency and explainability of AI decisions. If healthcare practitioners can understand how a decision was reached, reflecting on an AI system’s output becomes no different from any other diagnostic tool. If it can be proven that the duty of care was met, then the harm caused to a patient by an erroneous prediction of an AI-Health system would not constitute medical negligence. 510 However, it might also constitute negligence when healthcare providers fail to rely on the algorithmic output where the AI decision contains an obviously better treatment option for the patient. 511 Conclusion and recommendation This paper has attempted to show that although developing countries struggle with a high burden of disease, lack of trained healthcare providers, and poor healthcare delivery infrastructure, it is in these settings that AI has a tremendous potential to promote access to healthcare by reducing costs incurred due to accessing healthcare services, improving health equity, and improving the efficiency and quality of existing healthcare services. This technological advancement also improves existing healthcare systems, specifically in medical imaging and coronary artery disease diagnosis, by reducing human error, increasing patient care, and reducing the workload on healthcare professionals, which are currently reported as insufficient for the growing global populace. However, if the actual benefits of AI are to be gained, a collaborative approach should be encouraged between healthcare professionals and AI tools. As much as AI outperforms humans in data processing and analysis, human clinicians can exceed AI in the clinical decision making process, as human clinicians have direct interactions with their patients and access to clinical and contextual information. Also, the qualitative data collected through clinician intuition plays a critical role in clinical decision-making, thus ensuring the safety of patients. 512 The governments of developing nations should also be aware of the technological, ethical, and legal risks and challenges that arise when adopting AI, which must be addressed to ensure proper promotion of the right to health rather than perpetuating further harm. Meta-data generated from healthcare access, developed in the form of private sensitive and confidential information gathered in the process of healthcare delivery, is precious and priceless information that is required by the private companies who typically develop and run most AI tools, as such, securing this data is of vital importance and should be appropriately protected. Developing countries should be included in adopting this emerging technology as it has the potential to address many of the infrastructural deficits that currently plague 510 Holzinger A, Haibe-Kains B, Jurisica I, ‘Why Imaging Data Alone Is Not Enough: AI-Based Integration of Imaging, Omics, and Clinical Data’ (2019) 46 European Journal of Nuclear Medicine and Molecular Imaging 2722. 511 Schönberger D, ‘Artificial intelligence in healthcare: a critical analysis of the legal and ethical implications’ (2019) 27(2) Int. J. Law Info Technol. 171. 512 Chen A, Wang C and Zhang X, ‘Reflection on the Equitable Attribution of Responsibility for Artificial Intelligence-Assisted Diagnosis and Treatment Decisions’ (2023) 3 Intelligent Medicine 139.

120