Amazon’s facial recognition tool wrongly identified 28 U.S. lawmakers as people who have been previously arrested for a crime, the American Civil Liberties Union announced Wednesday.
The system, called Rekognition, created false matches that “were disproportionately of people of color,” according to the ACLU, which has previously criticized the technology for its surveillance capabilities.
The nonprofit tested the tool itself, as Amazon offers it to the public, and quite contentiously to law enforcement departments across the country.
“And running the entire test cost us $12.33 — less than a large pizza,” wrote Jacob Snow, technology and civil liberties attorney for the ACLU of Northern California, in a blog post Thursday.
The group was able to create a face database by browsing for 25,000 publicly available arrested photos and then searching for the available images of current elected officials.
The Congressional Black Caucus (CBC) — which includes mostly Democrats like Rep. John Lewis of Georgia, but also Republicans like Mia Love of Utah and Will Hurd of Texas — shared its apprehensiveness over Rekognition directly with Amazon CEO Jeff Bezos in a letter.
Due to law enforcement’s use of it, and in spite of claims that it’s allegedly being used to more efficiently identify criminals, “we are troubled by the profound negative unintended consequences this form of artificial intelligence could have for African Americans, undocumented immigrants, and protestors,” the CBC wrote on May 24.
“Our results validate this concern: Nearly 40 percent of Rekognition’s false matches in our test were of people of color, even though they make up only 20 percent of Congress,” said Snow.
Specifically, both the ACLU and CBC worry that there will be an increase in innocent people being accosted or confronted during investigations “based on a false identification.” (RELATED: DHS Seeking Facial Recognition Tech To Scan People’s Faces In Moving Cars)
“Face surveillance also threatens to chill First Amendment-protected activity like engaging in protest or practicing religion,” Snow continued. “Congress must take these threats seriously, hit the brakes, and enact a moratorium on law enforcement use of face recognition.”
Amazon touted its tool in a statement provided to The Daily Caller News Foundation:
We remain excited about how image and video analysis can be a driver for good in the world, including in the public sector and law enforcement,” said a spokeswoman for Amazon Web Services, its cloud computing department. “We think that the results could probably be improved by following best practices around setting the confidence thresholds. While 80% confidence is an acceptable threshold for photos of hot dogs, chairs, animals, or other social media use cases, it wouldn’t be appropriate for identifying individuals with a reasonable level of certainty. When using facial recognition for law enforcement activities, we guide customers to set a threshold of at least 95% or higher.
Send tips to email@example.com.
Content created by The Daily Caller News Foundation is available without charge to any eligible news publisher that can provide a large audience. For licensing opportunities of our original content, please contact firstname.lastname@example.org.