AI tool predicts biological age and cancer survival from facial photographs, study finds

Unbiased summary

Researchers at Mass General Brigham in Massachusetts developed an AI tool trained on over 40 million facial photographs to estimate biological age and predict cancer survival. Published in Lancet Digital Health, the study describes two complementary models: FAHR-FaceAge, which estimates biological age from features like nasolabial folds and forehead creases, and FAHR-FaceSurvival, which directly predicts survival risk from regions below the eyes and around the nasal bridge. The cancer estimates were validated against real patient data. The study found that cancer patients appear on average five years older biologically than their chronological age, and that lifestyle factors such as smoking, alcohol use, and drug use are associated with looking biologically older. The authors state this is the first demonstration that an AI model can directly predict cancer survival from a simple photograph.

Generated with deepseek-flash (DeepSeek)

Coverage by outlet

The Guardian left

Angle Frames the story as a significant scientific and technological breakthrough in AI health recognition, emphasising methodological detail and clinical potential.

Bias The Guardian provides the most detailed technical explanation, describing the two models (FAHR-FaceAge and FAHR-FaceSurvival) and the specific facial regions each analyses. It avoids sensationalist language, instead using measured terms like 'study suggests' and 'one of the most comprehensive demonstrations to date.' However, it downplays potential ethical concerns or limitations of facial analysis AI, such as privacy or bias, and does not include any critical perspectives. Its coverage remains largely factual but leans toward promoting the technology's promise.

New AI tool can predict cancer survival by a single patient selfie, study suggests

The Mirror centre-left

Angle Presents the story as a 'cancer breakthrough' with an accessible, consumer-friendly tone that highlights practical implications for readers.

Bias The Mirror uses sensationalist framing ('Cancer breakthrough') and emphasises the simplicity of a 'single selfie' to engage readers. It includes a promotional call-to-action for the outlet itself, which is unrelated to the news. While it reports key facts accurately, it omits technical details about the two distinct models and focuses instead on broad claims and quotes. It also highlights lifestyle factors like smoking and alcohol, potentially shifting focus toward individual behaviour rather than the technology's clinical application.

Cancer breakthrough as a single selfie could reveal your chance of surviving the deadly disease

The Independent centre-left

Angle Frames the story as a novel scientific finding with potential health insights, balancing accessibility with some methodological context.

Bias The Independent maintains a neutral, informative tone, similar to The Guardian but with less technical depth. It includes quotes from researchers and mentions the study's validation against real patient data, which supports credibility. However, it omits details about the two complementary models and the specific facial regions analysed. The article includes promotional elements (bookmarking, membership) but these are clearly separated from the news content. Overall, its coverage is largely factual and avoids sensationalism.

You can detect biological age and a person's chance of surviving cancer 'with just one photo'

The Sun right

Angle Presents the story as a health breakthrough with a practical, self-improvement angle, linking facial ageing to lifestyle and cancer survival.

Bias The Sun uses an enthusiastic tone ('A SINGLE photo... can detect') and emphasises lifestyle factors that readers can control, such as diet, smoking, and alcohol, with a sidebar on foods that accelerate ageing. It includes accurate technical details about facial regions but frames them in a way that suggests individual agency in health outcomes. The article omits any discussion of study limitations or ethical concerns about facial recognition AI. Its coverage leans toward sensationalism and lifestyle advice rather than a balanced report of the scientific findings.

Just one photo of your face could detect biological age and chance of surviving cancer