NYPL Letter — S.T.O.P. - The Surveillance Technology Oversight Project

December 15, 2020

Anthony W. Marx
President & Chief Executive Officer
The New York Public Library
Stephen A. Schwarzman Building
Fifth Avenue & 42nd Street
New York, NY 10018
Via U.S.P.S. & Email

Re: Face Detection Surveillance at New York Public Library Branches

Dear Mr. Marx:

My name is Albert Fox Cahn, and I serve as the Executive Director of the Surveillance Technology Oversight Project, a New York-based privacy and civil rights group. I was very disappointed and concerned to learn that The New York Public Library (“NYPL”) is using face detection and temperature scanning systems.[1] These types of surveillance devices are biased, invasive, and antithetical to the NYPL’s mission, and we ask you to cease such surveillance immediately.

Like many New Yorkers, the NYPL played a formative role in my own life, and I know you provide critical resources and safety net programming to so many New Yorkers. During the COVID-19 pandemic, it is more crucial than ever to ensure that the NYPL remains an open and welcome space for all New Yorkers. Sadly, thermal imaging kiosks are widely discredited as a public health measure. Not only do these systems fail to identify many COVID-19 patients and wrongly flag many others, they create significant risks of automated bias.

These kiosks are quite different from medical-grade temperature systems.[2] Kiosks often use skin temperature to determine internal body temperature, even though the two vary widely.[3] But skin temperatures fluctuate throughout the day and are impacted by environmental factors,[4] menstruation,[5] mere inflammation,[6] and many common medications.[7] Even if kiosks could reliably identify patients with a fever, they would do nothing to identify the many COVID-19 patients (perhaps a majority) who do not present with a fever.[8]

Face detection is biased, error prone, and antithetical to the inclusive values of the NYPL. Numerous studies have documented that facial detection systems’ error rates are highest for people of color and female-presenting individuals,[9] failing to even see these individuals as human beings.[10] This technology risks making the NYPL an invasive and segregated space,[11] directly undermining the NYPL’s mission.[12]

For decades, librarians have been some of our nation’s most vocal privacy advocates. NYPL staff have stood up to government censorship, opposed invasive surveillance laws, and spoken out in defense of the indispensable role of free inquiry in a democratic society. We ask you to reevaluate thermal imaging kiosks in light of your commitment to these values and to block any similarly invasive or biased surveillance in the future.

I would greatly appreciate if you could respond to this letter and provide the answers requested in Appendix A no later than Monday, January 4th, 2021. For the avoidance of any ambiguity, we have included images of the aforementioned face detection and temperature scanning system in Appendix B. If you have any questions, please do not hesitate to contact me at (646) 602-5602 or albert@stopspying.org.

Sincerely,

/s Albert Fox Cahn
_____________________________
Albert Fox Cahn, Esq.
Executive Director

CC:
Michele Coleman Mayes, Vice President, General Counsel & Secretary
Brian Bannon, Merryl and James Tisch Director of Branch Libraries and Education
Bill Marden, Director, Data Privacy & Compliance

[1] See Appendix B for photographs of these systems at the Harry Belafonte 115th Street Library.
[2] See Philippa Roxby, Can Thermal Cameras Help Spot Coronavirus?, BBC (June 9, 2020), https://www.bbc.com/news/health-52940951.
[3] Id.
[4] Ethan Ace et al., Detecting Coronavirus Fevers With Thermal Cameras, IPVM (Mar. 15, 2020), https://ipvm.com/reports/thermal-wuhan.
[5] Body Temperature: What Is (and Isn’t) Normal?, Cle. Clinic (Mar. 31, 2020), https://health.clevelandclinic.org/body-temperature-what-is-and-isnt-normal.
[6] Hannah Schaller et al., Thermal Imaging as Pandemic Exit Strategy: Limitations, Use Cases and Privacy Implications, Future of Privacy Forum (June 3, 2020), https://fpf.org/2020/06/03/thermal-imaging-as-pandemic-exit-strategy-limitations-use-cases-and-privacy-implications.
[7] John P. Cunha, Aspiring vs. Acetaminophen (Tylenol), RxList, https://www.rxlist.com/aspirin_vs_acetaminophen_tylenol/drugs-condition.htm (last visited Dec. 10, 2020).
[8] Interim Clinical Guidance for Management of Patients with Confirmed Coronavirus Disease (COVID-19), Ctrs. For Disease Control & Prevention, https://www.cdc.gov/coronavirus/2019-ncov/hcp/clinical-guidance-management-patients.html (last updated Dec. 8, 2020) (“In one study of 1,099 hospitalized patients, fever was present in only 44% at hospital admission . . . .”).
[9] See, e.g., Patrick Grother et al., Nat’l Inst. for Standards & Tech., NISTIR 8280, Face Recognition Vendor Test (FRVT) Part 3: Demographic Effects (2019), https://doi.org/10.6028/NIST.IR.8280; Joy Buolamwini & Timnit Gebru, Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification, 81 Proc. Machine Learning Res. 77 (2018), http://proceedings.mlr.press/v81/buolamwini18a/buolamwini18a.pdf.
[10] Larry Hardesty, Study Finds Gender and Skin-Type Bias in Commercial Artificial-Intelligence Systems, MIT News (Feb. 12, 2018), https://news.mit.edu/2018/study-finds-gender-skin-type-bias-artificial-intelligence-systems-0212; Rebecca Heilweil, Paranoia About Cheating Is Making Online Education Terrible for Everyone, Recode (May 4, 2020, 7:00 AM), https://www.vox.com/recode/2020/5/4/21241062/schools-cheating-proctorio-artificial-intelligence.
[11] See John B. Horrigan, Pew Res. Ctr., Libraries 2016, at 11–14 (2016), https://www.pewresearch.org/internet/wp-content/uploads/sites/9/2016/09/PI_2016.09.09_Libraries-2016_FINAL.pdf.
[12] NYPL’s Mission Statement, N.Y. Pub. Library, https://www.nypl.org/help/about-nypl/mission (last visited Dec. 10, 2020).