Thread by @CharlesRollet1

Scoop for @ipvideo: a facial recognition professor in Australia breached core ethics standards in a paper that improved Uyghur "ethnicity recognition" techniques, abruptly resigning

He now teaches AI at a top PRC university ipvm.com/reports/eth-re…

The study took pics of 300 Uyghur, Tibetan, and Korean PRC students to find the best 'face region' for 'ethnicity recognition'

The paper was published in 2018 as PRC police and surveillance companies ramped up 'Uyghur alarm' security camera systems across the country

The study sparked a @CurtinUni investigation into co-author and Curtin associate prof Wanquan Liu after it inadvertently went viral thanks to this @farbandish tweet

https://twitter.com/farbandish/status/1134485622352240640?s=19

@CurtinUni found the study "did not gain informed consent" or "required ethical approvals"

During the investigation, Liu "refused to respond" to certain questions - then abruptly resigned and now teaches at Sun Yat Sen University

Stunningly, the paper's publisher, Wiley, initially defended the paper saying face rec research "is enriched by expanding beyond white, Caucasian subjects" (the paper is explicitly about detecting repressed PRC ethnic groups, not reducing face rec bias) techinquiry.org/WileyResponse.…

@CurtinUni is now calling for the paper to be retracted and its name removed from all references. The paper has yet to be retracted; Wiley says it is "reviewing" this again

@IEEEorg has already retracted a separate paper based on the PRC minority faces dataset

@KU_Leuven professor Yves Moreau has been pushing for this paper to be retracted for years - he told @ipvideo facial recognition researchers have been ignoring basic ethics for far too long and are "playing with fire in an arsenal"

@_jack_poulson has also been on this case, emailing Wiley asking for answers and making replies public via @tech_inquiry

It took a while but the @CurtinUni findings and the @IEEEorg retraction are damning.

The ball is now firmly in @WileyGlobal's court.

this news comes as two DNA papers have been retracted for using Uyghurs' blood samples (collected by police!) with no evidence of consent, as reported by @nytimes' @suilee @paulmozur

https://twitter.com/paulmozur/status/1435975523360919564

Update: @SenPaterson tells @ABCnews it's "alarming to think an Australian university was involved with research that can so clearly be used for profoundly unethical purposes"
abc.net.au/news/2021-09-1…