# Can Artificial Intelligence Perform Health Assessments Using Selfies? Researchers explore using facial analysis AI to predict patient prognosis and detect medical conditions. - **Source:** [La Presse](https://www.lapresse.ca) - **Published:** 2026-11-07 - **Topic:** [health](https://brifing.si/c/health) - **Original article:** - **On Brifing:** https://brifing.si/s/821a57d4-7410-4898-9bc1-b7658aa741ef ## AI Facial Prognosis Research Harvard researchers recently published a study in The Lancet utilizing artificial intelligence to predict oncology patient mortality. By analyzing facial images taken via smartphones or computers, two distinct AI models evaluated age and mortality risks. This approach mirrors ancient medical observations regarding physical appearance as a diagnostic indicator for health. ## Mechanisms of Facial Analysis The AI models focus on specific facial regions to derive health insights. Age prediction relies on the nose, lips, forehead, and temples, while mortality prediction targets the area beneath the eyes and the upper nose. These findings align with Hippocratic descriptions of the 'mortuary face' observed in patients centuries ago. ## Clinical Application Timelines While current AI capabilities in medicine have evolved rapidly over the last three to five years, immediate clinical implementation remains pending. Experts suggest that while we are not yet at the stage of routine clinical use, significant transformations in diagnostic practices are expected to occur within the next decade of medical development. ## Blood Pressure Monitoring Challenges Researchers are investigating non-invasive blood pressure measurement using facial videos. Current challenges involve the necessity for weekly calibration against traditional arm cuffs. Experts like Robert Avram note that while wearable devices are being tested, achieving accurate, calibration-free readings remains a primary technical hurdle for widespread adoption in daily patient monitoring. ## Prioritizing Diagnostic Workflows AI integration in radiology aims to prioritize the analysis of urgent cases, such as strokes or pulmonary embolisms. By identifying subtle anomalies in medical images, algorithms help radiologists focus on the most concerning patient files first. This systematic triage could significantly improve response times and treatment outcomes in busy clinical environments. ## Predicting Cardiovascular Health A recent study in Nature Cardiovascular Research demonstrated that AI can predict arterial health from coronary angiograms with accuracy comparable to specialists. Trained on over 200,000 videos from 28,000 patients, the algorithm aims to reduce instances of incorrect treatment selection, ensuring patients receive appropriate interventions like stents or surgery. ## Ethical and Privacy Concerns The widespread use of facial data for medical diagnostics raises significant concerns regarding patient consent and data privacy. Experts worry about potential misuse by insurance companies or employers. There is a risk that sensitive health information derived from simple photographs could lead to discriminatory practices in life insurance or hiring processes. ## Future of Medical Training Concerns regarding 'cognitive abdication' suggest that over-reliance on AI might hinder the development of clinical skills in medical students. However, proponents argue that AI could actually enhance education by formalizing physiological observations that were previously considered intuitive skills possessed by only a small minority of exceptionally talented veteran physicians. ## Sources covering this story - https://www.lapresse.ca/actualites/sciences/2026-11-07/l-ia-peut-elle-faire-un-bilan-de-sante-a-partir-d-un-egoportrait.php?utm_source=rss&utm_medium=fil&utm_campaign=actualites — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._