July 27, 2018
Media Contact:
Rep. Jimmy Gomez – Eric Harris, 202-515-1301, Eric.Harris@mail.house.gov
Rep. John Lewis – Brenda Jones, 202-226-4673, Brenda.Jones@mail.house.gov
Rep. Tom Garrett – Andrew Griffin, 202-225-4711, Andrew.Griffin@mail.house.gov
ACLU – Abdullah Hasan, 646-905-8879, ahasan@aclu.org
Rep. Jimmy Gomez Leads Bipartisan Letter to Jeff Bezos Expressing Civil Rights Concerns about Amazon’s Facial Recognition Technology
Washington, D.C. – Today, Congressman Jimmy Gomez (D-CA) – joined by Rep. John Lewis (D-GA) and Rep. Tom Garrett (R-VA) – led a bipartisan letter to Amazon Chairman, President, and CEO Jeff Bezos, requesting a meeting about Amazon’s new facial recognition technology, branded and sold as “Amazon Rekognition.”
The letter, signed by 25 Members of Congress, expressed deep concerns about the civil rights implications such technology can have on constituents after the ACLU released the results of its own Rekognition test where 28 members of Congress were misidentified as suspected criminals. Among the matches were Rep. Gomez, Rep. Lewis, and Rep. Garrett. According to the ACLU, 40 percent of Rekognition’s false matches were of people of color, despite the fact that they make up 20 percent of Congress.
“The potential for this technology to wreak havoc on people of color cannot be overstated,” said Rep. Gomez. “Just one false identification in the field can cost our constituents their livelihoods, even their lives. It’s time for Mr. Bezos to recognize the role his company plays in this situation and to accept our bipartisan invitation to discuss these concerns and the impacts such technology can have on our constituents when put in the hands of law enforcement officials.”
“The results of the ACLU’s test of Amazon’s Rekognition software are deeply troubling. As a society, we need technology to help resolve human problems, not to add to the mountain of injustices presently facing of people of color in this country,” said Rep. Lewis. “It is not enough for Amazon to advise users of a 20 percent failure rate in their software. Law enforcement should not use this technology until the onerous civil rights and civil liberties issues are confronted and accuracy is guaranteed. If industry wants to engage in the public sphere, it needs to make the public good, not profit, a top priority. American families should not be collateral damage on the road to technological innovation.”
“This technology isn’t ready for prime time, and the risks to American’s basic liberties outweigh the benefits,” said Rep. Garrett. “I’m proud to stand alongside the ACLU and other champions of due process and civil liberties opposing the hurried imposition of potentially harmful technologies. That is why today, I have joined with my Congressional colleagues in calling on Jeffrey Bezos and the developers of Amazon Rekognition to confer with advocates and experts on civil liberties prior to introducing this technology into the marketplace or delivering it to law enforcement.”
“Members of both parties continue to sound the alarm about bias and accuracy issues with face recognition technology,” said ACLU Senior Legislative Counsel Neema Singh Guliani. “It is past time for Amazon to wake up and recognize that its product poses a grave threat to civil rights and civil liberties. The public demands action - not more excuses from the company that are aimed at hiding problems with the technology. Amazon should stop selling face surveillance to the government and Congress should issue a moratorium on law enforcement use until there can be a full debate on what - if any uses - are permissible.”
Full text of the letter can be found below, and a PDF can be found here.
July 27, 2018
Jeffrey Bezos
Chairman, President and CEO
Amazon.com, Inc.
410 Terry Avenue North
Seattle, Washington 98109
Dear Mr. Bezos,
We write to express our deep concerns about your facial recognition technology, sold and branded as “Amazon Recokgnition,” and the potential impact of this technology on the civil liberties, particularly in communities of color.
We request an immediate meeting with you to discuss how to address the defects of this technology in order to prevent inaccurate outcomes. Before Amazon Recokgnition or similar technology is on the market or used by law enforcement, we urge you to work with key stakeholders, communities of color, and policy makers on a regular basis, so that we can create a policy and regulatory environment that is able to adapt with the speed of technology.
As you know, a recent American Civil Liberties Union (ACLU) report found that your technology falsely matched 28 Members of Congress as suspected criminals. The ACLU discovered in its test, “nearly 40 percent of Amazon Rekognition’s false matches in our test were of people of color, even though they make up only 20 percent of Congress.”[1] Given the results of this test, we are alarmed about the deleterious impact this tool – if left unchecked without proper, consistent, and rigorous calibration – will have on communities of color; immigrants; protestors peaceably assembling and others petitioning the Government for a redress of grievances; or any other marginalized group.
Facial recognition and artificial intelligence technology represents how far technology has come and how far it has to go. While we understand that Amazon Rekongition is meant to “detect, analyze, and compare faces for a wide variety of uses,”[2] we are concerned that a match will exacerbate existing biases and strip people of their Constitutionally protected rights. There is an ongoing national debate about the critical need to address disparities in policing, sentencing, and the imbalance in the execution of justice in minority communities. In the hands of law enforcement, this technology would only aggravate this grave situation.
For these reasons, it is critical that developers of technology like Amazon Recokgnition should include testing and regular consultation with diverse stakeholders -- especially civil rights and civil liberty experts and advocates. Data software excels at finding patterns and aggregating data, but those programs have yet to prove they can accurately analyze the context around the information it collects. Without this context, facial recognition tools can falsely match individuals with those who have committed crimes while also exacerbating existing biases. This can and has led to irretrievable harm to innocent persons’ livelihoods; a cost that is too high to ignore.
Thank you for your consideration of our views on this grave matter. Technology should improve the lives and liberties of all people; it should not be a tool to oppress and suppress the rights of any person. We look forward to your prompt response.
###
Twitter: @RepJimmyGomez | Facebook: @RepJimmyGomez | Website: https://gomez.house.gov/
