Scientists New AI Tool Knows When You Will Die

By Robert Scucci | Published

Researchers from the Technical University of Denmark (DTU) have developed a new AI model called life2vec that can predict the future. Using a number of different input sources, including location of birth, education level, salary, and family health history, researchers have reason to believe that life2vec can map out an accurate sequence of future events based on what information is input.

Using past as prologue to make reasonable assumptions about future events, it’s not outside the realm of possibility that AI technology could one day accurately predict the time of one’s death with a relatively high degree of accuracy.

With a sample size of roughly six million people, researchers were able to create the AI model that could supposedly predict the future.

In a paper entitled Using sequences of life-events to predict human lives, researchers compare life’s linear timeline to the composition of a written sentence. In other words, like a sentence, human life also has a clear subject, form of action, beginning, middle, and end.

By using advanced AI learning models, researchers hope to one day be able to make realistic predictions that factor behavioral changes, possible future medical issues, and of course, human mortality.

ai

The study in question involved compiling health and labor data from Denmark’s population. With a sample size of roughly six million people, researchers were able to create the AI model that could supposedly predict the future. One thing worth noting, however, is that at the time of this writing there is no way to verify the research until further studies are conducted.

Ethically speaking, there are a lot of grey areas to consider.

On one hand, if AI could accurately predict future health issues that an individual may face, this could lead to breakthroughs in the medical sector. For example, if somebody is genetically predisposed to heart disease that will cut their life short, they could proactively make a series of lifestyle changes, or seek treatment that will allow them to extend their life.

AI models like life2vec could cause more harm than good.

Following this logic, more AI prompts would need to be processed based on these lifestyle changes, because positive changes in the present would most certainly affect the originally proposed future outcome.

Conversely, if insurance companies decide to take health and behavioral assessments generated by AI at face value, then they may not want to provide coverage to certain high-risk individuals. What’s more, employers could possibly think twice about hiring somebody to work for their company if they’re 35 years old, but the data suggests they’re only going to live until they’re 40.

Thinking in the most cynical terms, AI models like life2vec could cause more harm than good. Hypothetically speaking, if somebody has a history of substance abuse, and their proposed cause of death is related to their addiction, then they may never get access to the resources they need to course-correct, and live a long, fulfilling life.

By using advanced AI learning models, researchers hope to one day be able to make realistic predictions that factor behavioral changes, possible future medical issues, and of course, human mortality.

On the same token, if the same person has access to a highly accurate AI model that paints a bleak picture of their future, it may very well be the wake up call they need to reassess the aspects of their life that they could influence through making a series of positive behavioral changes.

At the end of the day, we’re still working within the realm of speculative science. In an ideal world, using AI to predict the future in this context could very well be the technology we need to carry humanity forward through future generations. But until we consider all of the implications, it’s probably in our best interest to tread lightly.

Source: Nature Computational Science