​A artificial intelligence neural network has identified a square structure in the triangular structure in the crater of the dwarf planet Ceres, and several researchers agree. This intriguing visual experiment results by Spanish neuropsychologists question the application of artificial intelligence in the search for extraterrestrial intelligence (SETI). Although Ceres is the largest celestial body in the main asteroid belt, it is a dwarf planet. A few years ago, it became famous for one of the craters:

Some highlights were observed there, which triggered various speculations. The mystery was solved when NASA Dawn probes discovered that these highlights were from emissions of volcanoes, ice and salt. Now, researchers from the University of Cadis in Spain, have studied one of the sites called, shocked by an area where geometric shapes can be seen on the surface. This feature led the researchers to propose an interesting experiment: comparing how humans and machines recognize planetary images.

The ultimate goal is to analyze whether artificial intelligence (AI) can help discover possible "technical features" of alien civilizations. Study author Gabriel G.de la Torre explains: "We are not fighting alone, and some people seem to be seeing a square, so we think it is a cognitive task in visual perception." In this regard, the opportunity to combat human intelligence and artificial intelligence is not only a routine task, but also a challenging task involving searches for alien life (SETI) rather than just radio waves.

The team of neuropsychologists from the University of Cadiz, who has studied undetected non-earth intelligent signaling problems, has now gathered 163 volunteers who have not received astronomy training to determine what they are in I saw something. The researchers then did the same with an artificial intelligence artificial vision system based on convolutional neural networks (CNNs), which was previously trained with thousands of images of squares and triangles to be able to recognize them. Both humans and AI detected a square structure in the image, but AI also identified a triangle when the triangle option was shown to humans:

The proportion of people who claim to see it also increased significantly, and this square appears to be embedded in the triangle. These results published in the Journal of Astronautics have led researchers to draw several conclusions: On the one hand, although artificial intelligence is popular and has many applications, artificial intelligence may confuse us and tell us that it detects it Something that is impossible or wrong. This impairs the effectiveness of AI in some cases in tasks such as searching for alien technology features, so care must be taken into account the implementation and use of AI in alien life (SETI).

On the other hand, if artificial intelligence identifies something that the human brain cannot understand or accept, will it surpass our level of consciousness in the future and open the door to reality that we are not prepared for? What if the squares and triangles of the light spot on Ceres were artificial structures? Artificial intelligence systems face the same problem as their creators: under human supervision, the impact of bias on their development should be further studied. But in reality, we don't know what it is, but what AI is detecting is likely to be just a game of light and shadow.

Boke Park｜Science, science, scientific research, popular science