[ad_1]
SARA ASHLEY O. BRIEN
NEW YORK – Democrat MP Jimmy Gomez was not too surprised to learn that he was one of only 28 members of Congress whose the faces coincided with police photographs in an experiment conducted by the United States Civil Liberties Union (ACLU) of Northern California. In fact, the California Democrat had a clear idea of what many people on this list had in common.
"I told my staff that I would not be surprised if it was mainly people of color or minorities". I was right:
Thursday, the ACLU released the results of its controversial software test Rekognition of Amazon, which uses, among other things, artificial intelligence for to recognize the people.
Lee: Rekognition, Amazon, a dilemma between security and the invasion of privacy
After building a database of 25,000 photos From public arrests and comparing them with the 535 members of Congress, the ACLU found that Rekognition – when left with its predetermined patterns – identified lawmakers such as Luis Guitierrez of Illinois, John Lewis of Georgia and Norma Torres of California, as criminals.
Of the 28 lawmakers that Rekognition falsely identified, around 40% were people of color, a result that led the ACLU to seek a government moratorium on the use of facial recognition software a call that gained strength among lawmakers on Thursday.
"It's a big problem," Gómez said. "The prejudice that exists will be digitized and will be used against people who already have a lot of obstacles and difficulties."
Although inaccuracy of facial recognition between women and men people of color is a known problem, experts point out the need for broader discussions about the ethics of this technology and the responsibility of the companies that create it to make sure that it is
to To this end, Gomez and Lewis have sent a letter to Amazon CEO, Jeff Bezos, asking for "an immediate meeting to discuss how to address the shortcomings of this technology in order to avoid inaccurate results".
In a separate letter, Senators Edward Markey, Mark DeSaulnier and Luis Gutiérrez asked Bezos to provide details of any internal badessment of precision or bias that Amazon made in Rekognition. They also want to know what the police or intelligence agencies use Rekognition, and if Amazon checks their use of the program.
Amazon criticized the results of the ACLU arguing, in effect, that the tool was not used properly
the results could probably be improved by following best practices to establish the confidence thresholds (that's the percentage probability that Rekognition finds a match) used in the test, "said an Amazon spokesperson in a statement.
80% confidence is a acceptable threshold for photos of hot dogs chairs, animals or other cases of use of social networks, it would not be appropriate to identify people with a reasonable level of certainty. we use facial recognition for police and judicial activities, we guide clients so that they establish a threshold of at least 95% or more. "
Read : Amazon goes to the plate … of your dog
Technology companies accept this vision. In a post earlier this month, Microsoft chairman Brad Smith suggested that facial recognition technology be regulated because of its "social ramifications and potential for abuse."
Bezos has not addressed publicly despite recent calls from shareholders and civil rights groups to stop selling technology to the government. In his statement to CNNMoney, Amazon called Rekognition "a good impeller in the world."
Perhaps it is, or maybe not, said Woodrow Hartzog, who teaches law and computer science at Northeastern University.
"The idea that it is simply a neutral technology that can be used for good or bad and that Amazon is not responsible, I think it's just wrong, "he said. a product that can damage, so you should be responsible for the design choices you make to allow damage, "he said," and when you launch it into the world, you do it in a safe and sustainable manner. "
We recommend: Jeff Bezos watch will give you the time of 10,000 years
[ad_2]
Source link
