What happened

PAHO hosted a session showing AI can help identify conditions and suggest codes from clinical data, but proposals require review and validation by health professionals. The process distinguishes AI-assisted coding from the final underlying cause-of-death determination, which remains under human responsibility.

Why it matters

The approach highlights mixed gains: faster, semantically aligned coding with quality checks, while preserving clinician oversight, ethics, and data governance.

PAHO's briefing describes AI as a practical aid for processing discharge or hospitalization information to flag relevant conditions and propose ICD-11 or ICD-10 codes. The final decision on the main condition and death sequence remains the clinician's responsibility, ensuring accountability and professional validation.

Limitations include the need for high-quality data, interoperability with standards, and transparent governance to prevent misinterpretation or overreliance on automated suggestions. Mixed gains depend on clinician oversight and data integrity, not on automation alone.

What this does not tell us

The piece relies on a single event description and publicly stated cautions about AI as a support tool; it does not provide independent validation of outcomes or broader population effects.

FOR PEOPLE

Benefits reported

AI helps but cannot replace a clinician's judgment.

FOR AI AND ITS OPERATORS

Benefits reported

AI supports coding with human oversight and governance.

These are two separate readings of what the sources describe. Reported claims and risks do not by themselves establish a real-world effect.

Original sources · 1
  1. Applications of Artificial Intelligence in the Use of ICD-11 for Morbidity and Mortality ↗Pan American Health Organization (PAHO) · 2026-09-03

Reporting discovered in Canada · United Kingdom. Discovery market does not mean the event happened there.