Exécution de notre outil de synthèse sur un nouvel article

Résumé : Age bias in medical AI is often treated as a performance problem, a perspective that fails to recognize an already existing risk: Generative systems may produce narrow, default representations of later life that enter health communication, patient personas, training materials, and care documentation. Because these representations potentially determine the needs, risks, and forms of support that take priority before clinical decisions are made, representational ageism should be addressed in bias audits, design documentation, procurement, and participatory review.