The study explores the ability of artificial intelligence to predict human life experiences

The study explores the ability of artificial intelligence to predict human life experiences

The researchers fed the AI ​​model Life2vec registration data for six million Danish citizens. The results that came back were amazingly accurate.

The rise of artificial intelligence has been made possible by large-scale language models – algorithms fed massive amounts of data to the point that they can understand and generate language at a speed and quality that is sometimes indistinguishable from a human.

But what if those predictive powers were applied to human life experiences? Can they predict the future?

In a new study, published Monday in the journal Nature Computational Science, researchers from the Technical University of Denmark and Northwestern University in Boston used artificial intelligence models to “examine the evolution and predictability of human life.”

To train the AI ​​model, known as Life2vec, the researchers fed it registration data for six million Danish citizens from 2008 to 2020, including a “daily analysis” of people’s education, residence, income, working hours and surrounding details. Their health, such as doctor and hospital visits, diagnosis, type of patient, and urgency of the visit.

Sonny Lehmann, a professor at the Technical University of Denmark and one of the study’s authors, said in a statement that the most exciting part of the study was access to the data that enables the AI ​​model “to provide such accurate answers.”

“The interesting thing is to consider human life as a long series of events, similar to how a sentence in a language consists of a series of words,” he said.

“We used the model to address the fundamental question: How well can we predict events in your future based on conditions and events in your past?”

The model predicted a number of diverse life outcomes, from subtle differences in personality to early death, with surprising accuracy.

The researchers focused on predicting people’s survival from 2016 to 2020, focusing on those aged 30 to 55 to make it more difficult.

The results were remarkably similar to what would be expected from studies in the social sciences, where people with higher incomes or in leadership positions were found to be more likely to live longer. Meanwhile, males, skilled workers and people with mental illness were found to have a greater risk of dying at an early age.

Diagrammatic representation of individual-level data for the Life2vec AI model. G. SAVCISENS et al.

The Life2vec model was also able to predict personality traits – such as introversion and extroversion – in line with the individual answers provided in the Danish Personality and Social Behavior Panel Study.

Researchers are now looking for ways to feed back typical images and information about social connections, Lehman said. The more precise the results are, the better researchers can “discover potential mechanisms that influence life outcomes” and find potential “personalized interventions,” the study said.

While many positive applications could show their findings, the researchers also offered a caveat: Their work was to see what was possible, and what should not, apply in the real world without regulations to protect individual rights.

You may also like...

Leave a Reply