WHO "Caution When Using Generative AI in Healthcare... Beware of Misinformation"

The World Health Organization (WHO) has warned that the use of generative artificial intelligence (AI), including ChatGPT, in the medical field may provide false information.

On the 16th (local time), WHO stated in a press release that strict verification and caution are necessary when using AI technology known as large language models (LLM).

LLMs process, analyze, and predict complex information based on knowledge obtained from massive data, enabling responses in natural language used by humans. This technology forms the basis of generative AI such as ChatGPT and Google Bard.

WHO expressed concern that "LLMs used without expert verification can lead to errors by healthcare professionals and harm patients," adding, "As a result, it can undermine trust in AI overall and damage the potential benefits of this technology." It further explained, "The data used to train AI in the medical field is biased, which can generate inaccurate or misleading information," and "It can be misused to create and spread false information that is difficult to distinguish from actual content."

Before widely using LLMs in everyday health management, pharmaceuticals, and other medical fields, WHO emphasized that policymakers and the medical community must address these concerns. While LLMs can have positive functions in healthcare, they must undergo strict verification and supervision.

Additionally, WHO stated that AI used in the health sector must be examined to ensure compliance with medical ethical principles. It presented principles including human welfare, safety, and public interest; transparent explainability and clarity; protection of autonomy; accountability; inclusiveness and equity; responsiveness; and sustainable AI.

This content was produced with the assistance of AI translation services.