How an ‘Artificial Intelligence Death Calculator’ Makes Its Predictions

AI can be used to predict pretty much anything, including life expectancy. Life2Vec‘s Artificial Intelligence Death Calculator enables users to interact with a chatbot and find out their life expectancy ‘with an accuracy per prediction between 70% and 90%’.

While it sounds like something straight out of the movie Minority Report, Life2Vec states that it provides a “fun” way for individuals to “assess their health and receive targeted health advice”.

In any case, the AI Death Calculator raises some big questions about whether AI can ethically be used to predict a user’s death. Let’s dive in.

Table of Contents Table of Contents

1. Key Takeaways
2. How the AI Death Calculator Works Under the Hood
3. Why The AI Death Calculator is Controversial
4. The Bottom Line
5. References

Show Full Guide

Table of Contents

1. Key Takeaways
2. How the AI Death Calculator Works Under the Hood
3. Why The AI Death Calculator is Controversial
Show Full Guide
4. The Bottom Line
5. References

How the AI Death Calculator Works Under the Hood

Life2Vec and the AI Death Calculator first gained popular attention in December 2023, when they released a paper titled “Using Sequences of Life Events to Predict Human Lives.”

This research analyzed the data of six million residents in Denmark for over a decade in an attempt to predict human life outcomes. But how does it work?

As part of the study, Life2Vec also collected a wide range of data about Danish citizens, including a labor dataset with records about income, salary, scholarships, job type, industry, social benefits, and other information.

This information was also considered alongside a health dataset, which included records about visits to healthcare providers, past diagnoses, patient type, and urgency.

The public AI Death Calculator takes a similar approach. Users input personal health metrics such as age, weight, height, daily calorie intake, and physical activity levels, and the solution then analyzes the data using machine learning algorithms.

More specifically, Life2Vec uses ‘long short-term memory‘ and recurrent neural networks to model sequential event data and then sequence-to-sequence architectures to map life event sequences to outcomes. This enables it to make reliable predictions about individuals’ life outcomes.

Why The AI Death Calculator is Controversial

But is there an ethical issue here? In the realm of this study and the calculator itself, researchers have been transparent about their research and the level of accuracy offered by the calculator, even admitting that it shows example outcomes and can’t act as a forecasting tool.

At the same time, predictions on life expectancy have been a staple in the insurance industry for decades.

Andrew Gamino-Cheong, CTO and co-founder at Trustible, told Techopedia:

“What they’re doing here is not fundamentally different than what insurance companies have been doing for decades.

“The whole field of actuarial science is just the old-school statistical version of this same task.

“The researchers have simply taken the same kinds of datasets and mapped them into a sequence structure compatible with neural networks.

“The biggest risk with these kinds of projects will be people confusing the correlation between a life event and their life expectancy prediction.

“The model they constructed can’t explain why certain events decrease expectancy, just that there is some degree of correlation there.”

Likewise, potential users who have personal or cultural beliefs that might dissuade them from using the AI Death Calculator (e.g., if they believe that predicting life expectancy can bring bad luck) have the option not to use it.

The Bottom Line

Above all, Life2Vec’s AI Death Calculator demonstrates how AI can be used to support a wide range of use cases, and that includes predicting life expectancy.

So long as users are aware of how it makes its decisions and use it to inform position health and lifestyle changes, it has the potential to be a force for good.

Insurance companies have been making educated guesses for a long time — this is just a way of democratizing the potential for the masses.