ChinAI #84: Biometric Recognition White Paper 2019
More from white papers on technical standards
|Jeffrey Ding||Mar 2|| 3|
Welcome to the ChinAI Newsletter!
Greetings from a land in which the residents seek to derive meaning and purpose amidst a stream of endless technocratic white papers…
…as always, the archive of all past issues is here and please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors).
Feature Translation: Biometric Recognition White Paper 2019
Context: Published in November 2019 by China Electronics Standardization Institute (CESI) & the Subcommittee on Biometric Recognition, National Information Technology Standardization (NITS) Technical Committee
Writing units included Sensetime, Xiaomi, Pingan, Ant Financial, Cloudwalk, Fudan University, etc. Notable absentees were Megvii and Yitu --- two of the “big four” facial recognition giants.
Sensetime’s leading role: it was appointed to lead a working group in charge of drawing up national standards for the facial recognition industry in December 2019. This is one of six working groups under the NITS Biometric Recognition Sub-Committtee, established in 2013. The other five mentioned in the White Paper are on iris recognition, vein recognition, behavior recognition, genome recognition, and mobile device biometric recognition standards.
In Sensetime’s Wechat post about the announcement, it claims to have been the chief compiler of 55 international, national, industry, and group standards, as well as obtained important seats in international standards organizations such as ISO and IEEE.
A lot of the big-picture implications about China’s push to shape AI standards was covered in an earlier DigiChina article I wrote with Paul Triolo and Samm Sacks.
The rapidly growing biometric recognition market in China: more than 4,000 companies in the field of biometric recognition, with 558 new companies in 2018. The size of China’s biometrics market has increased from 80 million RMB in 2002 to 17.01 billion RMB in 2018.
Government guidance has played an important role: “In 2012, Chinese investment and financing in biometrics was only 9 million RMB. With the government ’s policy support for biometrics in China, the scale of biometrics investment and financing reached 16.381 billion RMB by the end of 2018.”
Why does standardization matter: White Paper notes that currently there’s uneven product quality, inadequate ways to test/compare which algorithms are better. Section 5 of the report discusses CESI’s efforts to serve as a third-party testing agency to test whether biometric products comply with standards.
The Appendices reveal the systematic nature of this standards effort — Appendix 1 tracks 118 international standards in biometrics (along with which have been adopted domestically); A2 lays out the 43 biometric recognition standards published domestically; and A3 highlights 5 standards currently being developed. In 2019 the most recent one was on: “Information Technology: Biometric Recognition Presentation Attack Detection -- Part 1: Framework.” Note: think of presentation attack detection as anti-spoofing mechanisms. For instance, how to detect if someone is holding up an artifact sample (e.g. a printed photo) to try and fool the facial recognition system.
Blast from the past — let’s compare some of the stats from the White Paper to a 2006 presentation on China’s biometrics industry by a Professor at the Center for Biometrics and Security Research (part of Chinese Academy of Sciences and the largest biometrics team in China at the time ):
Gives a really nice way to track how the biometrics industry has evolved over the last decade or so: In 2006, fingerprint was 95.2% of market share, facial recognition comprised 1.1% of market share, and iris recognition was .5% of market share; in 2019, fingerprint recognition still leads with over 1/3 of the overall market share, but facial recognition is up to 16%, iris recognition is at 11%, and voiceprint recognition is at 11%.
China had just joined SC37 two years earlier in 2004 (The ISO/IEC joint committee branch focused on biometrics)
ChinAI Links (Four to Forward)
Must-read: China Digital Times Roundup on Concerns Over Adoption and Export of Biometric Surveillance
For more coverage on China’s use of biometrics in surveillance, see Samuel Wade’s excellent roundup in China Digital Times. For the other three links in this week’s Four to Forward, I’ll highlight a few from CDT’s roundup I found especially interesting.
From a FT piece by Yuan Yang and Madhumita Murgia on Chinese companies’ serving a growing international market, this paragraph caught my eye:
“US companies, for all the lip service they pay to technology and ethics, are also building surveillance tech, and indeed supplying Chinese companies that produce it,” says Stephanie Hare, an independent researcher of facial recognition technology and ethics, and a former employee of Palantir. “This leaves everyone else with a decision: be spied on by the US or by China? This point was made in the German parliament last week, and the US was very upset about it, saying there can be no moral equivalence between China’s authoritarianism and US values.”
By Sui-Lee Wee and Paul Mozur of NYT, this piece unpacks backlash to high-profile journals that publish papers by Chinese scientists affiliated with surveillance agencies — including a call for retraction of papers written by scientists backed by Chinese security agencies that focus on the DNA of minority ethnic groups.
Cai Xuejiao of Sixth Tone reports on a survey conducted by Nandu Personal Information Protection Research Center (a favorite of past ChinAI newsletters): “More than 73% said they would prefer alternatives to sharing their facial data, and 83% said they wanted a way to access or delete the data.” Nandu surveyed 6000+ people between October and November 2019.
Thank you for reading and engaging.
These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford, Researcher at GovAI/Future of Humanity Institute, and non-resident Research Fellow at the Center for Security and Emerging Technology.
Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).
Any suggestions or feedback? Let me know at email@example.com or on Twitter at @jjding99