Researchers have unlocked yet another artificial intelligence (AI) ability which, though intriguing, may not be as amusing as previous discoveries about the technology have been.
A recent study in Nature Computational Science suggests that a new AI system that treats human lives like language may be able to accurately predict death within a certain period, among other life details.
As part of the research, Danish researchers developed a machine-learning model called life2vec, which can predict people’s life details, including death, international moves, and personality traits.
The model uses data from millions of residents, including birth dates, sex, employment, location, and use of the country’s universal healthcare system.
Over four years, the model was found to be over 78% accurate in predicting mortality, outperforming other predictive methods like actuarial tables and machine-learning tools.
Life2vec showed promising early signs of connecting personality traits with life events, with a 73% accuracy rate in predicting people’s move out of Denmark and self-reported responses to a personality questionnaire in a separate test.
The study demonstrates an exciting new approach to predicting and analysing the trajectory of people’s life, says Matthew Salganik, a professor of sociology at Princeton University, who researches computational social science.
The life2vec developers “use a very different style that, as far as I know, no one has used before,” he says.
Lehmann and his team developed a language processing tool called life2vec, which can predict people’s future by processing individuals’ data into unique timelines of events like salary changes and hospitalisations.
The flexible model architecture allows for easy tweaking and fine-tuning offering predictions about many unexplored aspects of human life, making life2vec a promising tool for future prediction.
Lehmann says medical professionals have already contacted him to ask for help in developing health-related versions of life2vec—including one that could help illuminate population-level risk factors for rare diseases, for example.
He plans to use a tool to uncover hidden societal biases, such as unexpected connections between professional advancement and age or country of origin, and to explore the impact of relationships on quality of life and salary, as well as uncover hidden societal biases.