healthHealth and Medicinehealthhealth

AI Was Used To Guess Young Smokers' Age And The Results Were Brutal


Tom Hale

Tom is a writer in London with a Master's degree in Journalism whose editorial work covers anything from health and the environment to technology and archaeology.

Senior Journalist


An anti-smoking poster from the 20th century. Wellcome Collection CC BY

It’s not exactly news to hear that smoking is all-around terrible for you. But if you’re still looking for more inspiration to kick the habit in the new year, just remember that artificial intelligence (AI) is judging you.

Reporting in the journal Scientific Reports, researchers from US biotech company Insilico Medicine trained an AI algorithm to pick up on biological age differences between the blood of smokers and non-smokers.


In short, the AI thought that a young smoker was way older than their actual biological age.

“On average, female smokers were predicted to be twice as old as their chronological age as compared to non-smokers,” the scientists wrote in their research paper. “Male smokers, on average, were predicted to be one and a half times as old as their actual chronological age compared to nonsmokers.”

However, it's worth noting that this effect was only seen with smokers under 40 years old. The researchers go on to write: "Surprisingly, this effect disappears in the oldest subjects."

The researchers fed the AI a bank of blood profiles from 149,000 people, 49,000 of whom were smokers, gathered by routine blood tests. Using machine-learning techniques, it looked for patterns based on 66 different biomarkers in the blood that are known to relate aging, including levels of glycated hemoglobin, urea, fasting glucose, ferritin, etc.


After learning from all this data, the AI was then tasked to guess people’s age based on their blood biochemistry. They found that the machine-learning "aging clock" was fairly accurate at judging a healthy non-smokers' age. However, if a person was a smoker under the age of 40, it predicted their biological age to be "significantly higher than their chronological age."

Blood chemistry is just one of many ways to judge a person's biological age. While this is actually the first study to use blood test results to quantify smoking-related aging, other studies have looked at signs of aging, such as epigenetic aging, in relation to smoking. As you have probably guessed, those results were pretty damning on smoking too. 

"I am pleased to be part of the research study, which provides fascinating scientific evidence that smoking is likely to accelerate aging," Polina Mamoshina, a senior research scientist at Insilico Medicine, said in a statement“Smoking is a real problem that destroys people's health, causes premature deaths, and is often the cause of many serious diseases. We applied artificial intelligence to prove that smoking significantly increases your biological age.”


healthHealth and Medicinehealthhealth
  • tag
  • cancer,

  • tobacco,

  • aging,

  • smoking,

  • health,

  • age,

  • smoke,

  • cigarettes,

  • AI,

  • biological age