OAKLAND, CA – On Tuesday, July 16th, Oakland City Council President Kaplan’s Face Recognition ban was passed unanimously by the full Oakland City Council. (Link to Item)
Just as Oakland City Council President Kaplan introduced a ban on face recognition software in Oakland, the United Nations called for a moratorium on surveillance technology to end abuses. Said David Kaye, United Nations Special Rapporteur on the Promotion and Protection to Freedom of Opinion and Expression, “Surveillance tools can interfere with human rights, from the right to privacy and freedom of expression to rights of association and assembly.” (Link to Release)
A researcher with Microsoft described it as “toxic,” and called for it to “be banned for almost all practical purposes.” (Link to Article) In reality, Face Recognition systems rely on biased datasets with high levels of inaccuracy and lack standards around its use which has already lead to misidentification and manipulation of data. The invasive nature of this technology has also resulted in government abuses including its use to persecute Muslims in China and police accountability activists in Baltimore.
Oakland City Council President Rebecca Kaplan, an honors graduate of MIT, states: “I am thrilled my colleagues clearly understood the flaws in face recognition technology at this time. I welcome emerging technologies that improve our lives and facilitate city governance, but when multiple studies show a technology is flawed, biased, and is having unprecedented, chilling effects to our freedom of speech and religion, we have to take stand. It is important to build trust and good relationships between community and police and to remedy racial bias, however this flawed technology could make those problems worse. The right to privacy and the right to equal protection are fundamental and we cannot surrender them.”
Data shows that this technology disproportionately misidentifies darker skinned women. In a 2018 report by the MIT Media Lab, the study concluded that face recognition systems worked best on white males and failed most often with the combination of female and dark-skin individuals with error rates of up to 34.7%. (Link to Study) In another test by the ACLU, Amazon’s Rekognition face surveillance software misidentified 28 members of Congress as criminals. (Link to Test) The misuse and lack of guidelines around the use of this technology has also landed some police departments in hot water. In May 2019, Georgetown Law’s Center on Privacy and Technology (CPT) issued a report, detailing how some law enforcement agencies fed facial recognition software flawed data and warned that there are “no rules when it comes to what images police can submit to face recognition algorithms to generate investigative leads.” (Link to Report)
On May 2nd at the Oakland City Council Privacy Advisory Commission Chairperson Brian Hofer introduced the amendment that categorically prohibits the use of Facial Recognition technology. The amendment passed unanimously. (Link to Meeting Minutes – Item 6)