E-Paper
Can AI predict death? Here's what new study says
The study found that the AI model's predictions were 11 per cent more accurate than any other existing system or the methods used by life insurance companies
A new Artificial Intelligence model similar to ChatGPT can predict a person's time of death with a high level of accuracy. The model known as life2vec was trained on the personal data of Denmark's population. In light of a recent study, scientists at the Technical University of Denmark said that the model can predict the chances of a person dying more accurately than other systems.
In the study, which was published in the journal Nature Computational Science on Tuesday, researchers collected data from between 2008 and 2020 of 6 million Danes. It was used to analyse the health status and other information ranging from education level, doctor's appointments and hospital visits to resulting diagnoses, income, and occupation.
Scientists also took the data of people between the age group 35 to 65 to analyse the model's ability to predict life and death. Half of the data included those who died between 2016 and 2020. The study found that the AI model's predictions were 11 per cent more accurate than any other existing system or the methods used by life insurance companies.
DTU professor Sune Lehmann, who is also the first author of the article published in the journal said, “We used the model to address the fundamental question: to what extent can AI predict events in your future based on conditions and events in your past? Scientifically, what is exciting for us is not so much the prediction itself, but the aspects of data that enable the model to provide such precise answers.”
Lehmann added, “What’s exciting is to consider human life as a long sequence of events, similar to how a sentence in a language consists of a series of words. This is usually the type of task for which transformer models in AI are used, but in our experiments, we use them to analyze what we call life sequences, i.e., events that have happened in human life,” according to Independent.
The study also found that the model could predict general questions relating to the outcomes of a personality test, the chances of a person dying within four years, and more. However, the scientists warned that insurance companies should not use the system citing ethical reasons.
“Clearly, our model should not be used by an insurance company, because the whole idea of insurance is that, by sharing the lack of knowledge of who is going to be the unlucky person struck by some incident, or death, or losing your backpack, we can kind of share this burden,” Lehmann explained.
