Researchers in Denmark Create an AI Model to Predict Life

Danish researchers are using artificial intelligence (AI) in conjunction with vast amounts of data from many subjects to forecast various life phases, including death. Their initiative, life2vec, intends to highlight the potential power of this technology while simultaneously alerting users to its dangers. The goal of the study is to identify patterns that may predict different health or social events in an individual’s life using sophisticated computer systems.

According to Sune Lehmann, a professor at the Technical University of Denmark and research co-author, their system can forecast a lot of different aspects of people’s lives as long as there is data available. It may predict characteristics like financial success as well as health-related issues like obesity, cancer risk, or fertility. Though it focuses on analyzing life characteristics, including birth, education, social advantages, and employment habits, the system functions similarly to ChatGPT.

The technology behind language processing algorithms, such as those found in ChatGPT, is being used by the research team to analyze and forecast occurrences in human life. They consider a person’s life to consist of a sequence of events, including marriage, school, relocating, and birth. However, several individuals incorrectly referred to their application as a “death calculator” after hearing about it. Because of this misconception, phony websites have been tricking people into divulging personal information in return for a life expectancy estimate. According to the researchers, their program is not yet available to the general public or other researchers and is kept confidential.

Statistics Denmark provides anonymized data from around six million Danish citizens, which is used in the life2vec model. Based on an individual’s life story, this model may forecast various life outcomes, such as the likelihood of their death. It can forecast someone’s death with 78% accuracy and their migration to a different city or country with 73% accuracy.

The goal of the study is to forecast early death in adults between the ages of 35 and 65. To predict who will die in the next four years, they examine data spanning eight years (2008–2016). Lehmann claims that this model outpredicts mortality with more accuracy than any other algorithm they have tested.