A new artificial intelligence technology, Life2vec, has emerged, allowing each person to predict important stages in their lives, including death and levels of success.
This new technology explores the patterns that deep learning programs can detect, thereby predicting a range of conditions in social life. Trained on data from millions of people, this artificial intelligence program is working to provide these predictions.
“To predict human life, a very general model is needed,” said Sune Lehmann , a professor at the Technical University of Denmark and one of the study’s authors. This model can predict anything as long as there is enough data to train it.”
Life2vec’s lifespan prediction software algorithm uses a similar process to ChatGPT, but instead of analyzing text, it analyzes variables such as date of birth, education and social security.
Lehmann believes that the potential of this technology is endless. “It can predict health outcomes such as fertility or obesity. You can also predict whether someone will get cancer or not. This technology can bring huge benefits,” he said.
While this new technology is impressive, especially in predicting life expectancy, it also brings risks such as fraud. There may be fake and fraudulent websites advertising “death prediction” services to obtain users’ personal information.
However, the researchers claim that the software is protected by privacy and is not currently widely distributed to the public. “For now, this is just a research project, we are exploring its possibilities and limitations,” researcher Lehmann said.
With a prediction accuracy rate of over 73%, Life2vec’s algorithm was tested on anonymized data from around six million people in Denmark, collected by the country’s Official Statistics agency.
The results showed that the algorithm’s accuracy of this lifespan prediction software for predicting death was 78%. When predicting whether a person would move to another city or country, the results were correct 73% of the time.
“We were focused on predicting early mortality, so we selected a group of young people between 35 and 65 years old,” Lehmann explains. “We then tried to predict, based on an eight-year period from 2008 to 2016, whether they would die within the following four years.”
“Our algorithm was able to make this prediction, surpassing any other algorithm we tested,” he added.
According to the researchers, focusing on ages where mortality does not typically occur will help them verify the reliability of the algorithm.
