Blaze News investigates: Viral 'death calculator' AI tool cannot predict someone's time of death, but it could still be used to 'terraform society'

The creators of life2vec — an artificial intelligence tool that some claim can predict when someone will die — have warned that the viral coverage the tool has received is inaccurate. They noted that their creation cannot "predict the time of death, financial status, or health condition" of an individual.

The technology made headlines in December 2023, when U.S. and Danish researchers published a paper in the journal "Nature Computational Science" about the development of the AI tool. Since then, there have been claims the technology can accurately predict someone's death with 78% certainty. There have even been copycat AI technologies that claim to use the life2vec model, but the creators have said that their model is not yet available to the general public.

"The main 'advantage' provided by this technology, no matter how well-intentioned other aspects may be, is to more swiftly, effectively, and comprehensively terraform society into a more readily and fully manageable object[.]"

Therefore, the creators of life2vec have warned those interested in the technology to "be careful," adding that "we are not affiliated with these or any other entities that claim to use our technology."

The researchers pinpointed what the AI tool is designed to do, writing that "our model, life2vec, allows us to predict diverse outcomes ranging from early mortality to personality nuances. Using these methods for interpreting deep learning models, we probe the algorithm to understand the factors that enable our predictions. Our framework allows researchers to discover potential mechanisms that impact life outcomes as well as the associated possibilities for personalized interventions."

Despite the sophistication of the AI tool, there is no evidence to suggest that it can be used to predict precisely when someone will die. Instead, it uses AI technology to identify patterns in someone's life that may lead to early mortality or other health issues.

However, the development of life2vec — which, again, is still not available for use by the general public — raises questions about why and how it could be used in the future.

Brian Chau, executive director at Alliance for the Future, told Blaze News that "if this tool is able to accurately estimate the effect different life factors would have on life expectancy, it could help guide and prioritize which life factors you change in your life."

Chau went on to suggest that life2vec could function as a "diagnostic tool," which has been "a common part of medical treatment, and actuarial models have existed for centuries."

Actuarial modeling is generally used in the insurance industry. These models are made up of equations that represent the inner workings of insurance companies, which account for the probabilities of the events covered by policies and the costs each event might have on the company.

Dr. Kentaro Toyama — the W.K. Kellogg professor of community information and professor of information at the University of Michigan — appeared to share Chau's optimism about the tool, saying that "there are plenty of potentially positive uses of life2vec. For personal use, for example, a life2vec-based app could recommend customized changes to your daily routine to increase your life span, and it could conceivably do so without disclosing any estimates about when you might die."

"Similarly, policymakers could use it as a way to gauge the effect of a particular policy on a whole population," Toyama added.

While the researchers have noted that their AI tool is not able to predict death rates, they appear to be uniquely interested in the sequence of events that make up an individual's life. By using transformer models — which have been developed to find patterns in language — the researchers say they can "identify very complicated patterns in life-events."

"Just as in language, where the ordering of words is very important, so is the ordering of events in human lives. In a US context, for example, it matters if you get a job with health care and then get sick, rather than first getting sick without having the healthcare," the researchers stated.

Additionally, the researchers said they were interested in the subject of "predicting death" because "it is a problem so many people have worked on," such as insurance companies and those within the health care profession. They added they wish to be "in competition with many other algorithms" that exist within this same space.

James Poulos, host of the "Zero Hour with James Poulos" podcast, said the "uneasy reality" of this technology is that it "has as its primary use case a compulsory application."

"With death-rate prediction, the obvious use case is scaling the data so as to generate patterns, which can be applied to medical and 'care' industry efficiencies and reallocation of government resources to 'optimize' for 'end of life transitions.'"

Poulos added that "whoever exploits these goals stands to make a significant amount of money, but the motives here ultimately have to be understood as more spiritual than pecuniary."

However, it is widely accepted that AI is only as good and effective as the information fed into it. While there are healthy young people who could die tomorrow, there are also unhealthy chain-smokers who often live to old age. The complexity and variability of human life are not easy to package into information that can be fed into an AI tool.

"At some level, life2vec is only a more advanced version of the actuarial tables that health insurance companies already use," Toyama said. "So, I don't see that life2vec, or any similar predictive technology on the horizon, will lead to dramatically new forms of evil that the health industry isn't already routinely committing."

The creators of life2vec said that their overall goal is to use their algorithm to help "start a discussion about these technologies and how we should use them." They also mentioned that there is a reason why "Meta, Google, Microsoft, etc, collect so much data about us," adding that "predictions like these are already happening inside large tech companies."

It is well-known that the tech companies behind social media platforms use predictive algorithms to identify what users want to see on their digital feed, and the researchers for life2vec said they want to bring those closed-door predictive measures into public view.

Poulos concluded that the "main 'advantage' provided by this technology no matter how well-intentioned other aspects may be, is to more swiftly, effectively, and comprehensively terraform society into a more readily and fully manageable object — one that reduces human beings from the subjects we are into the objects we must, from the standpoint of the management system, become."

Like Blaze News? Bypass the censors, sign up for our newsletters, and get stories like this direct to your inbox. Sign up here!