Leveraging AI to Address the Alt Text Gap in Healthcare Websites

Web accessibility remains a significant challenge in the healthcare sector, where missing or inadequate alternative text (alt text) on images can impede access to critical information for individuals with visual impairments. This study investigates the potential of artificial intelligence (AI) to automate and enhance alt text generation on healthcare websites, with a particular focus on the role of contextual information in improving the relevance and quality of AI-generated descriptions. Ten major healthcare websites were analyzed, and images were sampled and processed using a context-free and a context-enhanced AI approach. Results indicate that AI-generated alt text, especially when supplemented with up to 1500 characters of surrounding webpage context, consistently outperformed both developerwritten and context-free AI alt text in terms of relevance, completeness, and alignment with accessibility standards. Context-aware AI was able to connect images to specific health topics, use appropriate medical terminology, and avoid redundancy, thereby improving both accessibility and user experience. However, the study also highlights the continued need for human oversight to ensure accuracy in complex cases. These findings suggest that integrating context-aware AI tools with editorial review can help healthcare organizations efficiently address accessibility gaps, comply with legal standards, and provide equitable information access. Limitations include the reliance on AI model capabilities and the need for further validation across broader website samples. Practical implications point to the feasibility of scalable, hybrid AI-human workflows for web accessibility, with significant social benefits for populations dependent on assistive technologies.