Scientists Train AI To Estimate How Much Time You Have Left To Live Based On One Photo Of Your Face

If you had only a few days left to live, would you want to know? Well, doctors do want to know, and they’ve now trained Artificial Intelligence (AI) to do the job.

Using nothing more than a single picture of your face, an AI tool can estimate your overall health and remaining lifespan. It can even tell you how old you look, as opposed to your actual age. While the technology is still in the works, experts say it could be a major breakthrough for treating cancer patients.

Let’s unwrap what this breakthrough means for all of us, how the tool was trained, and the serious privacy concerns it raises.

Scientists have built an AI tool that turns your selfie into a living medical blueprint

Igor Omilaev (not the actual photo)

A single photo can now reveal how well your body is actually aging

Developed by medical researchers at Mass General Brigham, a nonprofit in Massachusetts, this algorithm looks at a simple picture of your face to measure your overall health, your biological age, and how long you’re going to live for — basically, how well your body is holding up on the inside, regardless of how many birthdays you’ve had.

To teach the computer what to look for, scientists fed it a huge collection of more than 40 million photos — including 35,000 cancer patients — covering people of all ages, skin tones, and backgrounds.

The team ultimately built two separate tools that work together — FAHR-FaceAge and FAHR-FaceSurvival.

Zach M (not the actual photo)

Everyday habits leave subtle marks on your face

The two algorithms map your face in distinctly different ways to pull out these health insights.

To judge how fast you are aging, the FAHR-FaceAge tool zeros in on structural facial lines. It analyzes frown lines on the forehead, crow’s feet around the eyes, and cheek contours. It also examines the lines running from the nose to the mouth — known as nasolabial folds and marionette lines. The team discovered these specific areas clearly reflect how daily lifestyle choices, such as heavy drinking and smoking, accelerate the body’s internal clock.

When estimating immediate survival risk, FAHR-FaceSurvival completely ignores general wrinkles. Instead, it focuses heavily on the sensitive skin under your eyes and on the bridge of your nose.

“FAHR-FaceSurvival advances the field of facial health recognition AI a step further by learning survival risk directly from facial photographs of patients with cancer. To our knowledge, this is the first demonstration that an AI model can directly predict cancer survival from a simple photograph,” said Hugo Aerts, a co-senior author of the study.

According to preliminary results published in The Lancet Digital Health, a simple photo scan can give medical teams a quick way to check a patient’s physical strength. The goal isn’t to replace doctors, but to give them extra information. Knowing a patient’s true physical condition can help doctors tailor treatments and avoid overly harsh therapies

Glenn Carstens-Peters (not the actual photo)

How this AI tool works and what the findings reveal

Experts pulled most of the training photos from public websites like Wikipedia and IMDb, along with a massive picture database featuring people ranging from newborn babies to 116-year-olds.

The tool remained surprisingly consistent despite real-world variations in lighting, facial expressions, angles, and makeup. It even held up against most cosmetic surgery procedures, showing only a minor shift in readings after rhinoplasty (a nose job).

One of the study’s most striking results involved cancer patients receiving radiation treatment. Those who looked five or more years older than their actual age had a 21% higher chance of passing away over the next four years. Patients who looked at least five years younger than their age were 17% less likely to pass away during that same period.

Jud Mackrill (not the actual photo)

How human judgment stacks up against artificial intelligence

Researchers also wanted to see how much this new technology actually improves a doctor’s real-world decision-making. To test its value, they compared how accurately physicians could predict patient outcomes using photos alone, standard medical records, or a combination of both alongside the new algorithm.

“We found that doctors on average can predict life expectancy with an accuracy that’s only a little better than a coin flip,” when using a photo alone for their analysis, said Raymond Mak, a radiation oncologist at Mass General Brigham and one of the lead investigators for the study. 

Vitaly Gariev (not the actual photo)

Expanding the testing pool is the next step

Despite these promising results, researchers emphasize that the system is far from finished. To ensure the algorithms are accurate and unbiased for everyone, the team needs a much broader dataset containing images from diverse ethnic backgrounds.

That’s why they soon plan to launch a small pilot study with about 50 patients to see how well it works in real-life medical settings.

Getty Images / Unsplash (not the actual photo)

The AI tool also carries severe risks if it falls into the wrong hands

Unlike a blood test or a DNA sample, anyone can snap a photo of your face without you ever knowing. A single picture pulled from a passport scan or a social media profile could theoretically be used to gauge your health. For instance, wellness companies, life insurers, and employers could easily attempt to use facial scans to secretly evaluate how long you might live — potentially affecting your coverage or job prospects.

This technology can do a lot of good, but it could also potentially do some harm,” said Hugo Aerts. He added that hospitals have “very strong governance committees and regulatory guidelines that they have to adhere to, to make sure these AI technologies are being used in the right way, really only for the benefit of the patients.”

At the end of the day, this tech could be a massive game-changer for medicine. But the real plot twist depends on who controls it.

More info in the video below

Here’s what netizens think about this new AI technology:

Poll Question

Total votes ·

Thanks! Check out the results:

Total votes ·

Next question

Newsletter

Subscribe to Access
Exclusive Polls

Subscribe

By entering your email and clicking Subscribe, you're agreeing to let us send you customized marketing messages about us and our advertising partners. You are also agreeing to our Privacy Policy.

Thank you! You've successfully subscribed to newsletters!

Follow us on Flipboard.com/@boredpanda!