A global perspective on data powering responsible AI solutions in health applications

Area of Science:

• Health Informatics
• Artificial Intelligence in Medicine
• Global Health Equity

Background:

• Healthcare AI holds transformative potential for global access, quality of care, and patient outcomes.
• Current AI development often overlooks diverse global needs, particularly those of marginalized communities.
• A focused review on medical applications provides context for technologists facing development challenges.

Purpose of the Study:

• To review challenges in data, AI technology design, regulations, and infrastructure for global healthcare AI deployment.
• To highlight factors contributing to data gaps, regulatory hurdles, and connectivity issues.
• To emphasize the need for a global perspective in developing AI for healthcare.

Main Methods:

• Literature review focusing on medical AI applications.
• Analysis of challenges in data availability and quality.
• Examination of regulatory landscapes and infrastructural limitations (power, network).

Main Results:

• Significant gaps exist in healthcare data, particularly for marginalized populations.
• Regulatory frameworks and infrastructure (power, network connectivity) present major barriers to global AI deployment.
• Inadequate social systems for healthcare and education further impede universal AI impact.

Conclusions:

• Developing global healthcare AI requires a proactive approach to address data, regulatory, and infrastructural disparities.
• Prioritizing marginalized communities and a global perspective is crucial for equitable AI solutions.
• Recommendations are provided for developing prototype AI to better meet global population needs.