Researchers have developed an artificial intelligence system that can detect signs of life in soil samples from planets with up to 90 percent certainty. In combination with smart sensors on unmanned spacecraft, this system could contribute to our search for extraterrestrial life.
We are searching for life on other planets, but if we find it, will we know how to recognize it? To help us with this, American researchers have developed an artificial intelligence system that can tell whether substances in samples were made by living organisms or were created through other, non-biological processes.
Researchers from the Carnegie Institution in Washington, United States, analyzed 134 different carbon-rich samples to train and test the artificial intelligence system. Some of these samples had a biological origin, such as a grain of rice or a human hair. Other samples were abiotic and consisted, for example, of laboratory-made amino acids or carbon-rich meteorites.
Is there a risk in communicating with foreigners?
If there is intelligent life outside Earth, would it be a good idea to continue…
The AI was trained with measurement results from two-thirds of the samples. They used the remaining samples to test the system. Outside searchThe results, published in the journal PNAS, showed that the system correctly indicated whether a sample was biological or non-biological in 90 percent of cases.
Separate and define
The researchers analyzed the samples using A Gas chromatography mass spectrometry. This analytical device consists of a gas chromatograph, which can separate substances in a mixture, and a mass spectrometer, which can then identify the various components.
In gas chromatography, the sample passes through a narrow tube filled with a liquid polymer. The different components of the mixture have different chemical and physical properties, so they do not all pass through the tube at the same speed, so they are separated. Once they reach the mass spectrometer, the high-energy electrons destroy the separated materials. The parts measured by the detector provide information about the damaged material.
The researchers fed the results to the AI system and told it which graphs belonged to which biological samples and which did not. This information allows the system to determine whether unknown samples are of biological or non-biological origin. This is also called the strategy, through which the AI system is trained using data Machine learning.
Through this technology, we can know whether there is life or whether it existed in certain places. For example, we can examine very old samples from Earth to understand when life began on Earth. We may also be able to use samples from other planets to see if there are signs of life there.
The researchers trained the artificial intelligence system to distinguish between organic and inorganic carbon. To their surprise, the system can distinguish between not two, but three different groups: inorganic, living biological, and fossil biological.
It cannot be said with certainty whether an AI system can also recognize extraterrestrial life, which may be very different from terrestrial life. However, researchers are positive about this. They expect the alien non-biological samples to resemble non-biological terrestrial samples, which were very similar. Biological samples taken from another planet may differ from biological samples taken from Earth, but the materials they contain have been selected for their function in living systems. An AI system could potentially distinguish this from biological samples.
The system may eventually be able to determine specifically what type of life it relates to, for example, whether cells contain a cell nucleus and whether organisms are capable of photosynthesis. The system will also be able to determine how different life from another planet is from the life forms we know on Earth. This will allow us to know whether life on Earth and other planets came from a common origin, so-called panspermia, or whether they had different origins.
astronomer Leon Koopmans from the University of Groningen is convinced that the new artificial intelligence model is extremely powerful. “The big advantage of AI systems is that they are very fast at giving directions and can look for many patterns in a short time that humans would miss.”
Koopmans points out another important feature of the system. “Sometimes analyzes can also be performed on site, without having to first send soil sample measurement results back to Earth via a slow connection. In the case of a space or planetary mission, this saves a lot of time and energy.”
However, Koopmans believes we can rely on the power of artificial intelligence Machine learning It should not be exaggerated. “Currently, AI systems are only as good as the training data they rely on. The systems do not replace or produce the knowledge needed to train an AI system. This training information is often obtained through complex observations or laboratory work, for example This is by far the most work. I see AI more as a support than a replacement for humans.