Artificial intelligence has reached another milestone, as researchers unveil an unsettling tool that can predict the time of your demise with remarkable accuracy. Known as Life2vec, this AI death calculator utilizes transformer models, the same technology behind ChatGPT, to analyze the lives of individuals based on a sequence of events. By studying data from 6 million Danish people, the system can determine life expectancy by considering factors such as income, profession, and medical records. The results, published in the journal Nature, indicate that the predictions made by Life2vec have an astonishing 78% accuracy rate.
Sune Lehmann, one of the lead authors of the study, shed light on the workings of Life2vec, stating, We use the technology behind ChatGPT (something called transformer models) to analyze human lives by representing each person as the sequence of events that happens in their life.
To develop this life prediction model, researchers examined various aspects of individuals’ lives between 2008 and 2016. By seeking patterns within the collected data, they trained the algorithm to detect whether a person would have passed away by 2020. Notably, the Life2vec model achieved a high accuracy rate, providing researchers with intriguing insights into the influence of different elements on one’s lifespan.
The findings of the study highlighted several factors that contribute to longer or shorter lifespans. Individuals with higher income and those in leadership positions were more likely to enjoy a longer life. Conversely, harmful habits like smoking were found to reduce life expectancy. The AI death calculator also took into account additional factors such as exercise habits and mental health, providing a comprehensive analysis.
Encoded within the Life2vec model are various aspects representing different life events. For instance, a forearm break is symbolized by S52, while 072 denotes a postpartum hemorrhage. This encoding system offers a nuanced understanding of the events that shape an individual’s life and contribute to their predicted lifespan. Moreover, the model not only predicts when one might pass away but also estimates their financial situation at the time.
To test the accuracy of Life2vec, the system was deployed on a sample group of individuals aged 35 to 65 in Denmark, with roughly half of them already deceased. The model’s predictions proved to be remarkably reliable, making it a powerful tool for understanding mortality rates.
While the introduction of Life2vec may be unsettling for some, it offers valuable insights into the ways different aspects of our lives affect our lifespans. As with any new technology, it raises ethical questions regarding privacy and the potential misuse of personal data. However, the researchers behind Life2vec hope that their work will contribute to further advancements in understanding human longevity and ultimately lead to improvements in healthcare and quality of life.
As the field of artificial intelligence continues to evolve, we must grapple with both the immense benefits and unforeseen consequences it brings. Life2vec serves as a reminder of the extraordinary capabilities of AI and its potential to provide answers to questions we may not even want to ask.
