WHAT’S BEING CLAIMED:
- A new study by researchers from the University of Nottingham tested if an AI could predict the mortality risk of individuals.
- Using data collected from 500,000 individuals over the course of a decade, they used an algorithm to make the predictions.
- They found that machine-learned algorithms were even better at predicting when an individual would perish as compared to human-developed prediction standards.
Technology continues to progress in leaps and bounds, especially when it comes to AI. Well, we may be far from causing the next AI-robot-takeover, but the computational power that we have reached through the use of AIs is incredible.
Researchers from the University of Nottingham decided to test that power if an AI would be able to predict a person’s chances of mortality. The team used a huge amount of data gathered over the course of a decade from 2006 to 2016, covering a population of over a million people from the UK, with the ages 40-69.
Lead author of the work, Dr. Stephen Weng, said they developed a “unique and holistic approach to predicting a person’s risk of premature death by machine-learning.” This data included lifestyle habits, diet choices, and other demographic, biometric, and clinical data.
Predictions of the algorithm used in the study were mapped using “Office of National Statistics death records, the UK cancer registry and ‘hospital episodes’ statistics.” The results showed that machine-learned algorithms were even more accurate than standard models made by human experts when it came to predicting death.
The impressive study can be a big step in the right direction when it comes to creating machine learning systems that can customize health outlooks for different people. Perhaps in time, computers will be able to tell a person what illnesses they are at the highest risk for and how they can manage their lifestyle to avoid these health problems.