Technology company Alibaba has facial recognition technology that can identify members of China’s Uyghur minority, according to a new report on China’s artificial intelligence monitoring system released by the research organization IPVM.
China has repeatedly denied forcing anyone to enter what it described as vocational training centers, and also said: The Xinjiang region is threatened by Islamic militants.
Sensitivities have sparked caution among Chinese internet companies, which often practice self-censorship to avoid a confrontation with the government that strictly controls Internet speech, which last month published a draft of rules to monitor live broadcasts.
Research organization IPVM said: Software capable of identifying Uyghurs appears in Alibaba’s Cloud Shield Content Supervision Service for websites.
Cloud Shield is described as a system that detects and recognizes text, images, videos, and sounds that contain pornographic, political and terrorist material, ads and spam, and provides verification, tagging and other capabilities.
The archived record of the technology shows that it can perform various tasks, such as: checking glasses and detecting a smile, as well as knowing whether the targeted person is ethnically different, and whether he is a Uyghur.
According to IPVM, if a Uyghur broadcasts a clip live via a website registered with Cloud Shield, the system is able to detect that the user is Uyghur and report the video for review or removal.
“The mention of Uyghurs disappeared in the system after it published the report, and Alibaba told us that the feature was only used in the testing environment,” IPVM said.
Data from research market company Canalys showed that Alibaba – listed on the New York and Hong Kong stock exchanges – is the fourth largest cloud computing company in China and the fourth in the world.
US lawmakers earlier this month sent letters to Intel and Nvidia after reports of their computer chips were being used to monitor Uyghurs.