, 13 tweets, 7 min read
My Authors
Read all threads
The ethics bending research into the DNA of Uighurs we wrote about earlier this week is just the tip of the iceberg. A backlash questioning the ethics of a whole slew of studies, often seemingly targeting China’s ethnic minorities, has begun to emerge. nytimes.com/2019/12/04/bus…
One scientist reviewed 529 Chinese studies in forensic population genetics. He found a huge focus on ethnic minorities. Tibetans were 40 times more frequently studied than majority Han, Uighurs 30-times. Half the studies had some connection to police/military/justice officials.
That scientist, Yves Moreau, along w/ a group including @supernodal found problems in studies on biometrics and AI, and algorithms being developed to discriminate against groups like the Uighurs. We chronicled China’s automated racism earlier this year: nytimes.com/2019/04/14/tec…
@supernodal The group of scientists wrote to Wiley asking them to retract a paper that sought to perfect ways to identify Uighurs, Tibetans, and Koreans. At first the publisher rejected them. Now it is saying it has opened a new investigation. Here's the paper: onlinelibrary.wiley.com/doi/full/10.10…
@supernodal I found plenty of issues in Chinese biometrics studies. One 2016 paper called “Facial ethnicity classification with deep convolutional neural networks” used 100k+ images of Uighurs from China’s national ID database and surveillance cameras without any consent at all.
@supernodal Another study, seriously called “The detection of beard behavior of taxi drivers based on traffic surveillance video” seems in part based on surveillance footage taken from Beijing South Railway station. I say seems because they don't even specify, but the photos look like it.
@supernodal Other studies raised other questions. One analyzed cranial size of Uighurs to identify gender (without mentioning consent forms). Some assembled databases of ethnic minority facial expressions (Tibetan/Uighur), worked on ear tracking, iris maps, and detecting Uighur writing.
@supernodal Scientists point to conflicting standards. Many said there needs to be more guidelines for facial recognition papers that use people’s photos without their knowledge, whether from tech platforms or public surveillance. That’s a global problem. nytimes.com/interactive/20…
@supernodal Yet @supernodal points out if algorithms are being made to identify, and likely discriminate, against a particular minority, the subjects of the study should be made aware of that. In the Wiley paper they were not from sample unsigned consent forms we've seen.
@supernodal A few journals that act as lodestars for the scientific community are now reviewing their practices. Springer Nature is going through its entire catalogue, and said it found a number of problems. Wiley is reviewing another paper we mentioned. IEEE did not respond.
@supernodal If you’re interested in further reading, please look at Dr. Moreau’s paper published in Nature this week. It chronicles these issues, which may be severe in China, but are also present across the globals scientific community. nature.com/articles/d4158…
@supernodal With the power modern biology and computing coming together, it’s time to have a new think about standards. It's great some journals are carrying out these reviews, but more attention will be needed. Especially as China pushes the frontier on autocratic tech applications.
@supernodal Though many of the papers are less than cutting edge, this is the beginning of the new science for China’s surveillance state. Unsurprisingly the focus is hugely on minorities Beijing has long sought to control. nytimes.com/2019/05/22/wor…
Missing some Tweet in this thread? You can try to force a refresh.

Enjoying this thread?

Keep Current with Paul Mozur 孟建国

Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!