‘AI death calculator’ creators issue urgent warning about frighteningly accurate tool

Search for your death date — at your own peril.

Creators of Life2vec, the scarily accurate death calculator invented by US and Danish scientists, are warning of a new threat to those curious about learning when they kick the bucket — copycat “de-terminator” apps that hijack personal information.

“We are aware of Life2vec social media accounts, and there is at least one fraudulent website,” the creators warned while describing these morbid masqueraders, Metro reported.

For the uninitiated, Life2vec works by using ChatGPT technology to foretell when people will bite the dust based on select details from an individual’s life — including income, profession, residence and health history.

“We use the fact that in a certain sense, human lives share a similarity with language,” Sune Lehmann, lead author of a December 2023 study, told The Post. “Just like words follow each other in sentences, events follow each other in human lives.”

Most frighteningly, this digital prophet of doom can determine life expectancy correctly 78% of the time.

Unveiled in December, Life2vec is not yet available to the general public or corporations — the software is stored at Statistics Denmark — but that hasn’t stopped opportunists from trying to cash in on the macabre prognosticator.

The Life2vec team warned the public to “be careful of” bootleg websites imitating their creation and “have nothing to do with us and our work.”

“We are not affiliated with these or any other entities that claim to use our technology,” the creators declared.

These counterfeit de-terminators, of which there are a plethora online, often try to commandeer email addresses, phone numbers, credit card details and other sensitive info — they can even infect devices with malware.

It’s perhaps unsurprising that they’re trying to cash in on the deadly accurate tech, which along with predicting when people’s judgment day is coming, can also forecast if they’re going to be rich, among other things.

“This model can predict almost anything,” Lehmann, who teaches network and complex systems from the Technical University of Denmark, told The Post. “We predicted death because it’s something people have worked on for many years (for example, insurance companies) so we had a good sense of what was possible.”

To test out the AI’s algorithm, Lehmann’s team examined a heterogeneous subject population of 6 million Danish people, who varied in sex and age, between 2008 and 2020.

They determined which subjects would likely live for at least four years beyond Jan. 1, 2016, by feeding the AI-specific information to each study participant.

Sample data included “In September 2012, Francisco received 20,000 Danish kroner as a guard at a castle in Elsinore” or “During her third year at secondary boarding school, Hermione followed five elective classes.”

Each piece of data was assigned hyperspecific digital tokens — a forearm fracture was represented as S52, “postpartum hemorrhage” was O72 etc. — to calculate their approximate date of death.

Using said info, Life2vec was able to die-vine who had bit it by 2020 more than three-quarters of the time.

In other words, forecasting someone’s expiration date was not a roll of the “die.”

Of course, the team stressed that none of the study participants were provided their death predictions.

“That would be very irresponsible,” said Lehmann, who insisted that the purpose of the tech was not to satisfy morbid curiosity but rather to “understand what’s possible — and not possible — to predict.”