Scientists at the Technical University of Denmark (DTU) have developed an artificial intelligence model capable of predicting events in people’s lives, including death.
The technology, described in a study published in the journal Nature Computational Sciencehighlights both the power and potential dangers of this tool.
The model, dubbed life2vec, uses anonymous data from around six million Danes, provided by Denmark’s official statistics agency.
By analyzing sequences of life events such as birth, education, social benefits and even work schedules, the algorithm can predict a wide range of “life events”.
When it comes to predicting death, the algorithm is correct in 78% of cases; to predict whether a person will move to another city or country, the accuracy is 73%.
How AI works
The AI operates in a similar process to ChatGPT, but instead of analyzing words, it examines life-impacting variables. According to Sune Lehmann, professor at DTU and one of the study’s authors, life2vec is a “general framework for making predictions about human lives”.
It can predict health outcomes, like fertility or obesity, and even who might get cancer or make a lot of money.
Despite these successes, researchers are urging caution: the data only applies to Danes for now and therefore has built-in sociodemographic bias.
The team also wants to see privacy and personal data concerns addressed before this is used in a real-world scenario.
“This could be used against us to discriminate against us, so that you will have to pay higher insurance, or you won’t be able to get a loan from the bank, or you won’t be able to get public health care because you’re going to die anyway,” he told AFP Pernille Tranberg, Danish data ethics expert.
For now, the researchers explain that the software is private and unavailable on the internet or to the broader scientific community at this time.
){kind=link}