A neuropsychologist has warned against using artificial intelligence to detect possible signs of extraterrestrial life in images of distant planets and worlds.
Mainly because people will convince themselves that if an AI detects something, anything, there must be something to it.
What looked like a freaky face spotted on Mars by NASA's Viking 1 Orbiter in 1976 sent tabloid rags and conspiracy theorists into a frenzy. Later, though, higher resolution imagery showed it to be nothing of the sort.
And if you want to bring AI in to sort through photos and identify potentially alien structures, it better be good at detecting these structures – otherwise it'll just flag up blobs and and shapes to people who'll then convince themselves they've found E.T. That's according to research conducted by Gabriel De la Torre, an assistant professor of psychology at the University of Cádiz, Spain.
Torre conducted an experiment in which 163 volunteers and a convolutional neural network (CNN) studied a series of photos of the Occator crater on the dwarf planet Ceres, taken by the NASA space probe Dawn.
The humans identified what they thought were potential artificial alien structures, or so-called techno-signatures, in the blobs, squares, triangles, and other shapes on the crater's pockmarked surface. The CNN also flagged up shapes, because it was trained to do so, and in doing so, convinced the participants that something significant was found.
Crucially, De la Torre argued, if you use neural networks to search for alien life through thousands upon thousands, or millions upon millions, of input data samples, it must be properly trained, or you'll intensify people's biases by highlighting stuff they'll misinterpret.
Using AI to search for extraterrestrial life is, therefore, potentially useless as its observations could bias human investigators further, he argued, sending the public on wild goose chases. It seems obvious, though it's interesting and reassuring to verify in practice. This is science, after all.
Is it a square? Is it a triangle? Or ALIENS ... On the left, a shot of the crater, and on the right, structures people swore they saw (Source: De la Torre)
“We wanted to investigate how the search for techno-signatures or oddities might be influenced by our cognitive skills and consciousness, and whether artificial intelligence could help or not in this task,” De la Torre wrote in his paper published in Acta Astronautica.
“Both people and artificial intelligence detected a square structure in the images, but the AI also identified a triangle,” De la Torre added this week.
“And when the triangular option was shown to humans, the percentage of persons claiming to see it also increased significantly.
2001 fiction set to be science fact? NASA boffin mulls artificial intelligence to watch over the lunar GatewayREAD MORE
“On the one hand, despite being fashionable and having a multitude of applications, artificial intelligence could confuse us and tell us that it has detected impossible or false things and this, therefore, compromises its usefulness in tasks such as the search for extraterrestrial techno-signatures in some cases. We must be careful with its implementation and use in the search for extraterrestrial intelligence.”
If researchers train neural networks to look out for structures that seem odd to us humans then they will only carry the same biases that we have, he said.
“On the other hand if AI identifies something our mind cannot understand or accept, could it in the future go beyond our level of consciousness and open doors to reality for which we are not prepared? What if the square and triangle of Vinalia Faculae in Ceres were artificial structures?"
Fortunately, we don’t have to really take that question too seriously. A NASA study discovered that the bright spots in the Vinalia Faculae region nestled in the Occulus crater on Ceres are probably made out of highly reflective salt. We'll have to wait a little bit longer to find an alien species. ®