Health Equity at Risk? a Systematic Review of AI-Powered Health Education Apps and Emerging Disparities

Abstract

Background/Rationale
Artificial intelligence (AI) is transforming digital health education through mobile applications, offering scalable tools for information delivery and self-management. While these technologies have the potential to enhance health equity, they may also reinforce existing disparities in access, content design, and usability. This review critically examines AI-driven health education apps, identifying structural inequities and digital divides that limit their impact across diverse populations. It highlights the urgent need to ensure these tools serve, rather than exclude, marginalized and underserved groups.

Methods/Methodology
A systematic evaluation of AI-powered health education apps was conducted using the Mobile Application Rating Scale (MARS), assessing engagement, functionality, aesthetics, and informational quality. Readability was evaluated using the Flesch-Kincaid scale. Of 387 apps initially screened, 33 met inclusion criteria. Apps were categorized by target audience (e.g., youth, adults, specific health conditions), AI feature integration, pricing structure, and inclusivity in design. The analysis focused on how well apps addressed health literacy, affordability, and demographic-specific needs—especially those related to race, gender, and socio-economic status.

Results
Findings revealed stark inequities across the app ecosystem. Apps produced by government entities or major health organizations typically offered higher-quality, evidence-based content—but often featured complex language, limiting their usability among low-literacy populations. In contrast, apps from smaller developers had better readability but frequently lacked clinical accuracy or inclusivity. Only half of the apps provided fully free access, and one-third locked critical features behind paywalls, posing financial barriers for low-income users. Few apps demonstrated meaningful tailoring for diverse users, with limited consideration of racial, gender, or socioeconomic disparities in health education delivery. Moreover, while 17 apps incorporated AI-powered screening tools, most failed to apply these features in ways that supported equitable engagement or adaptive content personalization.

Conclusions
Although AI-driven health education apps offer innovative solutions for expanding access, this review underscores significant health equity challenges. Disparities in app design, content accessibility, and pricing models risk deepening the digital health divide. To truly advance health equity, developers must prioritize inclusive language, culturally responsive content, and affordability. Additionally, stronger regulatory and ethical frameworks are essential to guide equitable AI integration in digital health. Without deliberate attention to these challenges, AI risks reinforcing, rather than remedying, systemic barriers in health education.