Experts at the Washington University are designing a system to search the Earth for prehistoric animal bones with ease and with the aid of artificial intelligence, reports NewScientist.com.

According to Robert Anemone, co-developer of the project, finding fossils requires luck. Modern technology should change that and improve the chances of find old bones.

"We thought 'there's got to be a better way'," Newscientist.com quotes Anemone as saying. "We started talking about making a predictive model of where one might find fossils."

Their process involves using satellite data to map out encouraging fossil sites. The system would be fed with the criteria for a good site.

The first step involved the input of 10,000-square-kilometre region of the Great Basin in the Western U.S. The sites were either labelled as being fossil-rich, or were merely categorized as either barren, forest, scrub or wetland.

The system then categorizes the unknown location to fall under any of the five categories.

Two other criteria were programmed to improve the system's accuracy. First, is that the rock had to be around 50 million years old, the age of the fossils the team is studying. Second is that the land area must be sloping by at least 5 degrees, to raise the likelihood of the fossil being exposed due to erosion.

The research team then compared the predictions of their system with other known locations. The system identified 79 percent of the known fossil sites as prospective sites and correctly categorized 84 percent of all the other locations.

"[This would be] a more sophisticated way of doing something we've done for a long time", Peter Dodson, a palaeontologist at the University of Pennsylvania was quoted by Newscientiest.com as saying.

"Anything that helps palaeontologists is all to the good," he added.