As artificial intelligence improves, its applications are starting to show up in surprising places. For example, a new website from medical startup Insilico Medicine, Inc. wants to use AI to guess how old you are.
The site, Aging.ai, uses an artificially intelligent algorithm to analyze common blood markers like glucose and cholesterol.
They say that you can just enter these details from your latest blood test and the AI will try to guess your age and sex.
The company is trying to build AI that can make inferences about your health using deep neural networks modeled on the human brain.
Aging.Ai uses what we know about the changes in blood composition over time to guess your age. The idea is that if the guess is too far off from your actual age, you might start paying more attention to your overall health.
According to Alex Zhavoronkov, Insilico CEO and director of UK longevity think-tank Biogenontology Research Foundation, Aging.Ai has already been "trained on almost a million samples" to make inferences based on biomarkers.
The algorithm can guess sex, without using hormone levels, with 99% accuracy, and guess your age within a ten-year window at least 60% accuracy using ten key measurements found in a standard blood test.
With more information — there are up to 48 fields for your results — the accuracy rate increases to more than 70%.
"We want to incentivize people to be healthy," Zhavoronkov told Tech Insider. "As you try out various procedures, drugs, diets, or exercise, you may use Aging.Ai to see if it starts guessing a lower age."
Aging.aiDeep neural networks like the one in Aging.ai are a way to teach machines how to recognize patterns and make inferences based on those patterns.
If, for example, the computer sees that your cholesterol level is similar to the level it's seen previously with men in their 50's, that association gets taken into account, along with all the other pieces of information you've given it. All those considerations are then used for making a guess about your age and sex.
After it makes its guess, Aging.Ai will ask you how old you actually are in order to gauge the program's accuracy. The more people that use it, the more the algorithm will learn about how the markers relate to age and health, hopefully improving its guesses.
In general, research on exactly how the human body ages is still in its infancy, but promising. Blood markers especially also been a hot topic lately. Research suggests that blood tests can be used to flag your risk of dementia, osteoporosis, or cancer.
In recent years, there have been some promising advances hunt for biological markers that foreshadow how we age. If we understand the aging process, we might be able to predict or even slow the most drastic health effects of growing old.
Last April, researchers from the University of British Columbia found they could correctly determine the age of a tissue sample using chemical changes in DNA at the cellular level, which seems like a good first step. But there are still obstacles to the pursuit of a biomarker that has more predictive potential than DNA sequencing.
The American Federation for Aging Research notes that despite ongoing research, none of its criteria for an accurate and useful biomarker of "the underlying aging process," including the prediction of onset of age-related diseases, have been met.
When you use Aging.Ai, it won't tell you how long you'll live or what changes you need to make to stay healthy — Zhavoronkov hopes that your information will help improve AI recognize the signs of good health.
"We want robots to look at human lifespan as a game," he said — a game that artificial intelligence can help us win.