Evolution in the field of artificial intelligence seems to occur almost daily, with the latest update coming from a new study conducted by a team at University of Nottingham. The scientists tested three different types of artificial intelligence to predict people’s deaths. The most accurate model used a deep-learning algorithm, which was followed by the random forest model and the Cox model, noted Live Science.
The study drew from a large pool of data collected from 500,000 individuals over a ten-year span between 2006 and 2016. And during that time, about 14,500 individual passed away prematurely. The deep-learning model considered factors like air pollution, alcohol intake and use of different medications. It accurately predicted 76-percent of premature deaths. On the other hand, the random forest model considered skin tone, how much fruit and vegetables people ate, and body fat percentage. It predicted more or less 64-percent of premature deaths correctly. Finally, the Cox model focused on ethnicity and physical activity. This model predicted around 44-percent of premature deaths accurately.
There’s more to be done with the study, with researchers noting that it’s possible to further expand on the findings, according to Forbes. One of the scientists, Dr. Stephen Weng, noted that “We have taken a major step forward in this field by developing a unique and holistic approach to predicting a person’s risk of premature death, by machine learning.”
— AI News Feed (@AINewsFeed) March 29, 2019
Additionally, Weng expanded on what the findings of the study might mean for the healthcare industry.
“Preventative healthcare is a growing priority in the fight against serious diseases, so we have been working for a number of years to improve the accuracy of computerized health risk assessment in the general population. Most applications focus on a single disease area, but predicting death due to several different disease outcomes is highly complex, especially given environmental and individual factors that may affect them.”
Researchers seem focused on the ways that the new AI could be used to help patients if it’s used for preventative care. The co-author of the study, Joe Kai, also noted that their discoveries can be used to develop future methods.
It’s hard to know whether the AI technology would ever be available to the average consumer, or if industries will be the only ones with access. This also brings up questions about potential applications that could be invasive to privacy, like interpreted data being used to deny people’s applications to health insurance or life insurance policies.