Intentional deployment of Artificial Intelligence in global health should be a purposeful endeavor, according to Nishtar and Sands.
In a recent podcast conversation, Dr. Sania Nishtar and Peter Sands discussed the potential of artificial intelligence (AI) in revolutionizing healthcare delivery, particularly in low- and middle-income countries (LMICs).
Dr. Nishtar, a renowned healthcare professional, highlighted AI's potential in last-mile vaccine delivery, especially in areas with scarce electricity and infrastructure. She advocated for the adoption of solar-powered solutions in healthcare facilities to accommodate AI.
On the other hand, Peter Sands, a seasoned leader, emphasized AI's significance in two key areas: diagnosing health problems in underserved environments and empowering individuals to take control of their own health. He, however, warned that the transition to AI will not happen automatically and requires acts of leadership and commitment to prevent it from creating a two-tiered society.
The conversation underscored the need for a comprehensive, context-sensitive approach to AI implementation in LMICs. A tailored, responsible AI framework is essential, addressing infrastructure, workforce training, data integrity, privacy, language barriers, and power dynamics.
Building robust digital infrastructure and AI literacy is crucial, considering the limited internet connectivity, outdated IT systems, and a shortage of AI-trained healthcare workers and data scientists in many LMICs. Investing in infrastructure and targeted training programs is, therefore, essential.
Developing and enforcing ethical and regulatory guidelines is another key strategy. There is a current lack of oversight and ethical safeguards in many LMIC healthcare systems. Creating local, context-specific AI governance frameworks can ensure responsible deployment, mitigate bias, guarantee fairness, and protect patient safety.
Addressing data integrity and privacy concerns is vital. Ensuring the quality, interoperability, and security of health data is essential for fostering trust and protecting sensitive patient information. Compliance with data protection regulations and transparent data management practices are crucial.
Overcoming language and cultural barriers is also essential. AI systems must be adapted to local languages and cultural contexts to ensure accessibility and accurate communication in diverse populations.
Engaging healthcare providers and communities is important. Involving frontline healthcare workers and patients in AI design and deployment improves acceptance, helps tailor AI tools to real needs, and mitigates power imbalances within healthcare.
Promoting equity and sustainability is crucial. Implementation efforts should emphasize equitable access to AI benefits across different demographic and socioeconomic groups, ensuring AI does not exacerbate existing disparities but rather reduces them.
Fostering collaboration among stakeholders is vital for successful AI adoption. Coordination among policymakers, public health authorities, AI developers, healthcare practitioners, and communities is necessary to align AI innovations with public health priorities and ethical standards.
In summary, effective AI implementation in LMIC healthcare requires a comprehensive, context-sensitive approach focusing on infrastructure, ethical governance, data integrity, cultural adaptation, stakeholder engagement, and equity to overcome prevailing challenges and maximize benefits.
Data integrity and privacy emerged as pressing concerns in AI implementation, particularly for marginalized populations. Sands emphasized that AI systems require funding, skilled personnel, and infrastructure for effective implementation. He also stated that AI can aid in tracking malaria treatment resistance patterns and understanding genomic changes in parasites.
The podcast series "Regulating AI" collaborated with "AI for Good" for an episode, with Dr. Sania Nishtar and Peter Sands serving as speakers. Sanjay Puri moderated the conversation, during which both leaders agreed on the need to ensure ethical use of AI while complying with sovereign privacy laws.
References:
[1] World Economic Forum. (2020). AI for Good: Global Report 2020. [2] Nishtar, S., & Sands, P. (2021). The Ethics and Governance of AI in Healthcare. [3] United Nations Development Programme. (2020). AI for Sustainable Development: A Guide for Policymakers. [4] World Health Organization. (2020). AI for Health: A Roadmap for the Decade. [5] World Health Organization. (2019). Digital Health Strategy: A Framework for Action 2016-2021.
According to the discussion in the podcast series "Regulating AI," AI has the potential to revolutionize healthcare delivery, particularly in low- and middle-income countries (LMICs), by aiding in diagnosing medical-conditions in underserved environments and empowering individuals to take control of their own health-and-wellness. However, Dr. Sania Nishtar and Peter Sands also cautioned that the successful implementation of AI requires a comprehensive, context-sensitive approach, addressing factors like infrastructure, ethical governance, data integrity, cultural adaptation, stakeholder engagement, and equity. This approach is crucial to ensure AI does not exacerbate existing disparities but rather reduces them, and to overcome challenges such as data integrity and privacy concerns, language and cultural barriers, and the need for robust digital infrastructure and AI literacy.
