Overview
- SARAH, the WHO's AI-powered health chatbot, often provides incorrect or outdated medical information.
- Despite being available in eight languages and designed to offer empathetic responses, SARAH's effectiveness is limited by its outdated training data.
- The chatbot is part of the WHO's efforts to address global health worker shortages but lacks diagnostic capabilities and often refers users to health professionals.
- Researchers and the WHO acknowledge SARAH is a work in progress, emphasizing the need for improvements in accuracy and data security.
- Concerns about privacy and data security persist, as SARAH uses facial recognition and open-source data, making it susceptible to cyber threats.