An algorithm to know the day of your death

When it comes to predicting futures, artificial intelligence and specifically algorithms are used to prevent mortality and also to point out future crimes, but…the date of our death? This is what scientists from the Technical University of Copenhagen (DTU) would have developed, led by Sune Lehmann, professor of Networks and Complexity Sciences.

According to Lehmann's team, artificial intelligence developed to model written language can be used to predict events in people's lives. The project shows, according to the university statement, that if large amounts of data about people's lives are used and so-called “transformative models” are trained, which (like ChatGPT) are used to process language, they can systematically organize the data and predict what will happen in a person's life and even estimate the time of death.

“Our models allow us to predict various outcomes that range from early mortality to personality nuancessurpassing the most modern models by a wide margin,” point out the authors of the study published in Nature.

To reach this conclusion, Lehmann's team analyzed health and labor market engagement data from 6 million Danes in a model called life2vec, which acts as a “life events calculator.” After training the model in an initial phase, that is, learn data patternshas been shown to outperform other advanced neural networks (see data box) and predict outcomes such as personality and time of death with high accuracy.

“We use the model to address the fundamental question: to what extent can we predict events in your future based on conditions and events in your past? Scientifically, what is exciting for us is not so much the prediction itself, but the data aspects that allow the model to provide such precise answers,” says Sune Lehmann, in the statement.

Life2vec predictions are answers to general questions. When researchers analyze model responses, the results are consistent with existing findings in the social sciences; For example, all things being equal, people in a leadership position or with high incomes are more likely to survivewhile being a man or having a mental diagnosis is associated with an increased risk of dying. Life2vec encodes data into a large vector system, a mathematical structure that organizes different data. The model decides where to place data on time of birth, schooling, education, salary, housing and health.

“The interesting thing is to consider human life as a long sequence of events, similar to how a phrase in a language consists of a series of words. This is normally the type of task that transformative models are used for in AI, but in our experiments with them we analyze what we call life sequencesthat is, events that have occurred in human life,” adds Lehmann.

However, the authors point out that their model is surrounded by Ethical issues, such as the protection of confidential data, privacy and the role of bias in data. These challenges must be understood more deeply before the model can be used, for example, to assess an individual's risk of contracting a disease or other preventable life events.

“The model opens up important positive and negative perspectives to discuss and address politically – concludes Lehmann -. Similar technologies for predicting life events and human behavior are already used today within technology companies that, for example, They track our behavior on social networks, they profile us with extreme precision and use these profiles to predict our behavior and influence us. “This debate must be part of the democratic conversation so that we consider where technology is taking us and whether this is the development we want.”

The next step would be to incorporate other types of information, such as text and images or information about our social connections. This data use opens a completely new interaction between social and health sciences.