AI 'death calculator' can predict when you die - study

An AI-powered 'death calculator' claims to predict the day you die with disturbing accuracy with just four data points.

Death calculator (photo credit: INGIMAGE)
Death calculator
(photo credit: INGIMAGE)

An artificial intelligence-based "death calculator" chatbot known as Life2vec claims to be able to predict the date a person will die with just four data points, boasting a 78% accuracy rate, a recent study revealed.

This may represent a groundbreaking advancement in disease prevention and life expectancy extension and could have the potential to revolutionize the way we approach personal health.

Can AI tell you when you'll die?

Unlike other models of "death calculators," Life2vec stands out by functioning as a chatbot like ChatGPT. Developed by a team of Danish and US scientists, the algorithm was trained on a database of six million Danish citizens.

The inputs required by Life2vec include just four personal data points:

  1. Income
  2. What you do for work
  3. Where you live
  4. History of pregnancy and injury
 An illustrative image of an artificial intelligence (AI) bot. (credit: INGIMAGE)
An illustrative image of an artificial intelligence (AI) bot. (credit: INGIMAGE)

To test the accuracy of the calculator, the team analyzed a group of people between the ages of 35 and 65 from 2008 to 2020. Half of the subjects studied had died between 2016 and 2020.

The study found that Life2vec's predictions were 11% more accurate than any other existing AI model.

"The whole story of a human life, in a way, can also be thought of as a giant long sentence of the many things that can happen to a person," explained Sune Lehmann, the lead author of the study from the Technical University of Denmark (DTU). 

Lehmann stressed the potential of this calculator to help individuals change their lifestyle and reduce the likelihood of premature death. "We can... learn from [life2vec] what the factors are that might help you live longer," he noted. 

Published in the journal Nature Computational Science.

Advertisement