Facial recognition fail: Amazon project mistakes lawmakers with suspects
Amazon’s facial recognition tool wrongly identified 28 members of Congress as police suspects, disproportionately matching lawmakers of color with the mugshots, according to the ACLU.
The American Civil Liberties Union (ACLU) used Amazon’s powerful image ID software, Rekognition, to compare photos of members of Congress to a database of 25,000 police suspects.
We used Amazon’s facial recognition tool to compare photos of members of Congress to a database of mugshots — we got 28 false matches.And even though they only make up 20% of Congress, nearly 40% of the false matches in our test were members of color. https://t.co/WdNRWtqZfa
— ACLU (@ACLU) July 26, 2018
Republican and Democratic legislators of all ages, men and women, were among the mismatches. However, people of color were disproportionately wrongly matched, according to the ACLU, even though they make up just one-fifth of Congress.
These instances included six members of the Congressional Black Caucus, among them Democrat politician and civil rights leader John Lewis. In May, the Congressional Black Caucus wrote to Amazon CEO Jeff Bezos to raise concerns over the “profound negative unintended consequences” face surveillance could have for black people, undocumented immigrants, and protesters.
The ACLU is now calling on Congress to back its campaign for a moratorium on law enforcement’s use of the technology, citing its worrying experiment. “People of color are already disproportionately harmed by police practices, and it’s easy to see how Rekognition could exacerbate that,” it warned.
The organization said it paid just $12.33 to have Amazon Rekognition compare official photos of every member of the US House and Senate against a database of 25,000 public arrest photos, using the default match settings.
It first raised concerns about Amazon’s marketing of the service as a government surveillance tool in May. Bezos’ corporation claimed that deployment by law enforcement agencies was a “common use case” for the technology and documents revealed that police departments in Orlando and Oregon’s Washington County were already using the real-time facial recognition system.
READ MORE: Insane Clown Privacy: Juggalo makeup is the key to defeating facial recognition software
Rekognition allows one to scan mug shot photos against real-time footage such as surveillance or body cameras. The tool has also been touted as an effective means of detecting offensive content online and identifying celebrities.
“We remain excited about how image and video analysis can be a driver for good in the world,” a spokeswoman for Amazon Web Services said in a statement to Reuters, pointing to how it can be used to find lost children and prevent crimes. She added that Rekognition was normally used to narrow the field for human review, and not to make final decisions.
Responding to Ars Technica about the ACLU experiment, a company representative claimed that the non-profit should have set a high confidence threshold for its test.
"While 80 percent 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," the company said. "When using facial recognition for law enforcement activities, we guide customers to set a higher threshold of at least 95 or higher."
If you like this story, share it with a friend!
