Artificial Intelligence and Health Misinformation in Healthcare: Opportunities, Limitations and Future Directions

The rapid digitalisation of health information has transformed how individuals access medical knowledge but has also facilitated the widespread dissemination of health misinformation. The World Health Organization describes this phenomenon as an infodemic, characterised by an overabundance of information that includes both accurate and misleading content, making it difficult for individuals to identify reliable sources of health advice. Exposure to misinformation has been linked to vaccine hesitancy, reduced adherence to public health guidance and increased engagement with unverified treatments. Artificial intelligence (AI) technologies are increasingly being explored as tools to detect, monitor and counter misleading health narratives across digital platforms. This article critically explores current evidence on AI-based approaches to health misinformation governance, including machine learning, natural language processing, deep learning, large language models, multimodal systems and social media surveillance. Evidence suggests that AI can help identify misleading narratives and emerging public concerns, but current studies rarely distinguish intentional disinformation from unintentional misinformation when reporting model performance. Significant challenges include dataset bias, restricted generalisability across linguistic and cultural contexts, platform-specific model development, limited external validation, privacy concerns, uncertain clinical relevance and amplification through engagement-driven recommendation systems. The article argues that AI should be understood as one component within an iterative, risk-stratified governance framework that combines healthcare expertise, public health interpretation, transparent oversight, platform accountability, tailored digital health literacy interventions and evaluation of real-world public health outcomes.