AI death calculator accurately predicts when you’ll die

While many avoid knowing their end, a new AI death calculator eerily predicts precise demise.

“We use the technology behind ChatGPT to analyze human lives by representing each person as the sequence of events that happens in their life,” said Sune Lehmann, lead author of the December 2023 study ‘Using sequence of life-events to predict human lives’, to The New York Post.

In the study, the professor of network and complex systems at the Technical University of Denmark, along with co-authors, presents an algorithm called "life2vec." This algorithm utilizes specific details of a person's life, such as income, occupation, location, and health history, to predict life expectancy with an accuracy of 78%.

“We use the fact that in a certain sense, human lives share a similarity with language,” said Lehmann. “Just like words follow each other in sentences, events follow each other in human lives.”

Differing from ChatGPT, the constantly buzzing bot embraced by tech experts for securing dream jobs or crafting ideal outfits, life2vec calculates life outcomes by closely analyzing the individual histories of men and women.

“This model can predict almost anything,” Lehmann told The Post, he also shared that his research team also used the specialized program to foretell people’s personalities and decisions to make international moves.

“We predicted death because it’s something people have worked on for many years (for example, insurance companies),” he added.

Lehmann's team studied a diverse population of 6 million Danes of different ages and genders from 2008 to 2020. Using life2vec, they identified individuals likely to survive at least four years beyond January 1, 2016, as per the reports.

“The scale of our dataset allows us to construct sequence-level representations of individual human life trajectories, which detail how each person moves through time,” states the report.

“We can observe how individual lives evolve in a space of diverse event types”, he added.

The researchers provided the AI with detailed information about each participant in straightforward language, like "In September 2012, Francisco earned 20,000 Danish kroner as a guard at a castle in Elsinore" or "During her third year at secondary boarding school, Hermione took five elective classes."

Subsequently, they assigned distinct digital tokens to categorize each piece of data specifically. For example, a forearm fracture was denoted as S52; working in a tobacco shop was coded as IND4726; income was represented by 100 unique digital tokens, and "postpartum hemorrhage" was coded as O72, as reports suggest.

With the given information, life2vec accurately predicted deaths by 2020 with an accuracy of over 75%.

According to the research, factors such as gender, mental health diagnoses and skilled professions may contribute to earlier death. Conversely, higher income and leadership roles were associated with a longer life.

Lehmann told The NY Post that participants were not informed of their death predictions.

The bot is presently inaccessible to the general public and corporations. If it eventually becomes available for widespread use, the researcher suggests it probably won't be utilized to inform individuals directly, for example, in writing insurance policies or making hiring decision.