Artificial intelligence as a factor of freedom and speed, but also of greater control over the masses. In China, a country where there is one surveillance camera for every two people, and artificial intelligence is sometimes embedded in connected devices to track people.
First modification:
By Louise May and Stéphane Lagarde, RFI correspondents in Beijing
Dialogues with artificial intelligence begin in the morning in Beijing. Here, a conversational speaker: “Xiaodu, can you tell me what the weather is like today?” Answer: “It’s a little less cold than yesterday, with light rain and 30 degrees on the thermometer. “Our hyper-connected lifestyle, with access to transportation, our homes and electronic payments, means that artificial intelligence is producing mountains of data in China, both images and sounds.
Smile, they’re listening! RFI was at Security China 2023, the international exhibition for Chinese public security professionals. One company told us about surveillance microphones in public places.
“In fact, it works by keywords and voice identification. If the system hears someone fall to the ground or someone calls the police in a crowd, it immediately transmits the information to the control center. The same happens if it hears ‘thief’, for example. At the moment, the system only works with conversations in Mandarin,” the company representative told RFI.
Tracking of individuals lost in the crowd
The sales representative for this maker of surveillance tools did not tell us about other keywords, sensitive terms that can be added for more political surveillance. Artificial intelligence built into the public safety apparatus’s facial recognition cameras, controls and cloud makes it possible to track individuals embedded in the mass. It is the so-called “one person, one file” system, developed since 2016 by dozens of Chinese companies present at the security fair and who collaborate closely with the Government.
Artificial intelligence can also be used to improve hygiene, says another exhibitor. “This device is designed above all to monitor schools and nursing homes. First, images are captured with cameras and then algorithms are used to identify anomalous situations. That’s it. That person in the kitchen has not covered their hair! Someone is smoking in the corner. And here passes a cat, a dog or a rat”.
Chinese authorities concerned about ChatGPT
Artificial intelligence is on alert, alarms go off or, on the contrary, when faced with certain questions, the robots remain silent. “What happened on June 4, 1989?” [matanza de la plaza de Tiananmen de Pekín]. She says that she doesn’t know. Chinese chatbot amnesia follows the contours of official censorship. But that does not apply to ChatGPT, whose extreme freedom worries the Chinese authorities. Yu Yang is an adjunct professor in the Department of Artificial Intelligence at the prestigious Tsinghua University.
“Personally, I am in favor of not letting ChatGPT into China, because it would be a problem for national data security. More generally, all AI algorithms should be audited before commercial use. Always check their social impact So we need audits for all forms of artificial intelligence used in public companies, but also for the private sector, particularly the services sector.”
This systematic audit is mentioned in a bill unveiled last month. Artificial intelligence tools must reflect “core socialist values,” the Chinese president claimed, while calling for greater Chinese state control over AI. “We must prepare for the worst… to be able to withstand strong winds, rough waters and dangerous storms,” Xi Jinping was quoted as saying by the China News agency.
The announcement of stricter controls also has a deterrent effect. The Chinese don’t know if they are being watched, and the uncertainty makes them more compliant, the system’s designers believe.