PAHO publishes guide for designing Artificial Intelligence instructions in public health – PAHO/WHO

Washington DC, October 20, 2025 (PAHO) – The Pan American Health Organization (PAHO) has launched a new guide that provides practical advice on how to create Artificial Intelligence (AI) messages that generate reliable, relevant and culturally appropriate content in various contexts. The publication aims to promote the responsible use of AI to improve communication and decision-making in public health.

With the advent of generative AI, technology has advanced from information analysis to content creation. Public health institutions are increasingly using AI to write alerts, translate reports into plain language, develop educational materials, and simulate responses. Tasks that previously required multiple rounds of review can now be completed more efficiently, as long as the AI ​​is guided by clear, specific, and purposeful instructions.

“Generative artificial intelligence has become a powerful tool for public health, but its effectiveness depends on how it is instructed,” said Marcelo D’Agostino, Head of the Information Systems and Digital Health Unit at PAHO. «Good rapid design is key to unlocking its full potential,» he added.

The publication, Rapid AI Design for Public Healthunderlines that the value of generative AI depends on the ability of professionals to effectively guide it to produce accurate and meaningful content.

In its simplest form, a message refers to the instructions given to an AI system, whether it is a short question like “How can dengue fever be prevented?” or a more detailed request specifying tone, format and audience, for example: “Write an educational message about vaccination for parents in rural areas, in clear and friendly language.”

For this reason, designing effective messages can be considered an essential skill for all public health workers: improving operational efficiency and ensuring that messages are reliable, understandable and actionable. However, the guidance also warns of the risks associated with using generative AI in public health, particularly when the content could influence public behavior, inform local policies, or support emergency response efforts. Emphasizes the importance of human oversight to review and approve all AI-generated results.

Among its recommendations, PAHO suggests that instructions be treated as “life protocols”: instructions that can be tested, refined and adapted depending on the context, language or audience. It also encourages institutions to develop rapid libraries to promote consistency and efficiency in the use of AI across public health operations.

This publication is part of the PAHO Digital Literacy Program, which aims to strengthen the digital competencies of public health professionals in the Americas. It also supports the digital transformation of health systems, allowing faster, more accurate and higher-impact decision-making to improve people’s health.