Artificial intelligence has tracked down 11 new killer asteroids that could smash into Earth and level cities.
Astronomers claim a supercomputer managed to locate asteroids NASA was unable to analyse.
It then predicted how the space rocks would travel through the solar system over the next 10,000 years.
A from Leiden University in the Netherlands developed the tech, dubbed the Hazardous Object Identifier, to hoover up space data missed by NASA.
Researchers say the newly reclassified space rocks were in a "chaotic" orbit – making their trajectories difficult to plot. Although they were known to NASA they were not deemed dangerous until now.
Each is believed to over 100m across, with the potential explosive force of hundreds of nuclear weapons, which could decimate whole regions of Earth if they hit.
Asteroid killed dinosaurs after hitting Earth with force of 10 billion nukes
Giant flying dinosaur which could travel across world dug up in Australia
The scientists, who published findings in the journal Astronomy & Astrophysics in February, predict one of the killer rocks could slam into Earth from 2131.
Leiden University said in a statement the killer astroids "have the potential of causing regional devastation, unprecedented in human history" if they were to slam into Earth.
It continued: "That these asteroids have not previously been identified as potentially dangerous is because the orbit of these asteroids is so chaotic.
Psychic who predicted JFK assassination claims 'the world will end in 2020'
What is an asteroid? One space rock can be worth £4 TRILLION
"As a result, they are not noticed by the current software from space organisations, which is based on probability calculations that use expensive brute force simulations."
Read More
Today's Top Stories
Brit couple catch coronavirus on cruise Football coach raped and murdered HSBC to axe 35,000 jobs 'Danger to life' flood warnings
The team is now working on making its neural network even more accurate.
“We now know that our method works, but we would certainly like to delve deeper into the research with a better neural network and with more input,” co-author Zwart said.
Source: Read Full Article