IPVM, a video surveillance research company, recently discovered a Huawei document describing an interoperability test on a technology that triggers a ‘Uighur alarm.’ This feature notifies government authorities whenever a Uighur is identified.

According to the Washington Post, John Honovich, the founder of IPVM, said the document showed how “terrifying” and “totally normalized” such technology has become: “This is not one isolated company. This is systematic,” Honovich said. “A lot of thought went into making sure this ‘Uighur alarm’ works.”

This document was shared exclusively with the Post and has since been taken off Huawei’s website.

On the left is the Huawei logo. On the right is Megvii Face++ software that was displayed during an expo. Source: NY Post

What the Technology Entails

The artificial intelligence camera system is a combination of Huawei hardware and Megvii’s deep learning and image recognition software. The tool is made to scan faces in a crowd and estimate each person’s age, sex and ethnicity.

This usage itself is controversial because, in the past, facial recognition technology has proved to be especially inaccurate in detecting minority groups, resulting in unfair treatment. Since minorities have been discriminated against using facial recognition in the past, implementing a technology that further widens the inequality gap is counterintuitive.

However, the most concerning part of this technology is by far the ‘Uighur alarm.’ The camera system is trained to trigger an alarm if a member of the mostly Muslim minority group is seen. This alert could potentially be used to flag the individual for the police as a part of the Holocaust-esque crackdown on Uighurs.

Since the release of these documents, Huawei and Megvii executives have announced that they are not directly involved with the imprisonment or abuse of Uighurs. Yet, the document said that the Huawei/Megvii system could pull 10 seconds of video before a Uighur is identified and 10 seconds after. This was considered a “basic function.”

The camera system also includes a language recognition feature. In light of the fact that Uighur is a unique language, spoken predominantly in Xinjiang, this feature could also be used to identify Uighurs.

The Response

Huawei and Megvii have both recognized the document as authentic since the Washington Post’s story was released. In response to the news, a Megvii spokesperson said the company’s systems are not designed to target or label ethnic groups. A Huawei spokesperson also said that what the document describes “is simply a test and it has not seen real-world application” and that “Huawei only supplies general-purpose products for this kind of testing.”

While the companies’ responses show that they are not actively working to create discriminatory technology, they also don’t show a concern for the potential negative applications of this technology. With the right financial incentives, could technology like this be produced? Furthermore, could Chinese officials demand this type of technology in the name of safety?

Only time will tell the answers to these questions.

ABOUT THE WRITER

Sarah Raza is a Tech Innovation Fellow with a background in computer science from Stanford. She is passionate about exploring the implications of increased usage of artificial intelligence and machine learning.

Contact Sarah Raza at sarah@identityreview.com.

Do you have information to share with Identity Review? Email us at press@identityreview.com.