Evidence of an ancient fire was revealed via an artificial intelligence tool at an archaeological site in Israel, which is about one million years old. If the AI technology is also used on other sites, it could help in revolutionizing the understanding of where and when human beings had learnt how to control fire. After all, the ability of making and controlling fire is considered as one of the most notable achievements to have been made by human beings.
Techniques for identifying use of fire
There are already a couple of techniques that are used by archaeologists for identifying whether fire was used by ancient humans, or not. For instance, discolored prehistoric bones are one sign and warped stone tools are also considered another indication. This should be in a manner that is consistent when temperatures go above 450 degree Celsius. However, finding this kind of evidence is tough when the site is about 500,000 years old, or even older.
A deep-learning AI tool had been unveiled by a group of researchers in Israel last year and this artificial intelligence tool had the capability of identifying subtler indications of fire that occurs due to exposure to temperatures between 200 and 300 degree Celsius. The algorithm had been trained by the team at non-archaeological sites in the countryside to gather bits of flint. These had then been heated to specific temperatures in the lab.
The flint would obviously respond differently to UV light and the AI had been given the task of identifying the changes in response, no matter how subtle they may be.
1 million year old fire identified
Hailing from the Canadian University of Toronto, Michael Chazan worked with the team and the algorithm was used for assessing the flints gathered from a human site in Israel called Evron Quarry. This is about one million years old. Chazan stated that they had decided to use this site because it has the same kind of flint used previously. However, they did not know that they would find evidence of fire.
Therefore, it came as a surprise when the artificial intelligence tool identified that a lot of the flint tools at Evron Quarry had been heated and exposed to temperatures as high as 400 degree Celsius. The team decided to use existing techniques for testing the chunks of bones found at the site and got the same results. According to Chazan, they would not have bothered to test them for heat if the AI had not given the flint results.
The evidence at Evron Quarry is not of natural wildfire because of how the heated bones and stones are clustered. There is little evidence at this point that human beings were using fire 1.5 million years ago, but Chazan said that the AI tool can come in handy at other sites to find more evidence. He said that they can use artificial intelligence to discover if fire was in use by humans, for cooking and otherwise, even 2 million years ago.