Researchers from Wuhan University conducted an experiment with the active participation of artificial intelligence in space. According to the SCMP, the Qimingxing 1 Earth remote sensing satellite was given to the AI for a day.
The article says that artificial intelligence independently, without the intervention of people, chose the observation targets that were interesting for itself.
The first thing the AI did was point the cameras at the ancient city of Patna along the Ganges River in northeast India. This place is known for the base of the Bihar Regiment. In 2020, he was involved in a clash with Chinese army units in a disputed border area.
Another high-priority area on the list of targets for AI surveillance was the Japanese port of Osaka, where US Navy ships call from time to time.
Note that scientists were worried about the fact that the operation of observation satellites is often ineffective. For example, China now has more than 260 Earth remote sensing satellites, but scientists say they often work “idlely”, collecting data of no practical value. In order for the collected data to be as relevant as possible for certain clients, according to the researchers, it is necessary to create a trainable model.
For this task, they assembled a large language model like ChatGPT and endowed it with the ability to take the initiative. As an experiment, she was given a remote sensing satellite at her disposal for a day, allowing her to independently choose the objectives of the observation. The developers believe that AI should understand the relationship of events not only based on learning from databases, but also on the example of direct observation of nature and human activities.
One of the Chinese scientists, who was not involved in the study, said he was not worried about the prospect of giving AI more control over the satellites. “I believe that the Beijing Mission Control will keep a close eye on everything,” he said, adding that if the AI tries to do something “wrong”, it will be immediately interrupted by a human operator.
However, another expert with experience with ChatGPT expressed concern about what might happen if a large language model gains access to flight control and learns to control the satellite. “Satellites are our eyes and ears in the sky. Do we really want AI to decide what we see and hear?” he said.
Source: Rosbalt
Mario Twitchell is an accomplished author and journalist, known for his insightful and thought-provoking writing on a wide range of topics including general and opinion. He currently works as a writer at 247 news agency, where he has established himself as a respected voice in the industry.
