Department of Homeland SecurityDocuments obtained through an information release issued by the Government Oversight Project (POGO) show that the Amazon…
Department of Homeland Security
Documents obtained through an information release issued by the Government Oversight Project (POGO) show that the Amazon Government’s sales unit was actively trying to provide Department of Homeland Security Immigration and Customs Enforcement (ICE) Homeland Security Surveys with Recognition, the Controversial Facial Recognition System. The pitch was part of a larger discussion about Amazon Web Services offers to ICE HSI, an offer including artificial intelligence algorithms and predictive analysis.
Amazon has provided cloud services to DHS earlier, including US Citizenship and Immigration Services (USCIS). USCIS uses Amazon Cloud Storage to store data, including information associated with Alien Registration Numbers (A-Numbers) numbers also used by ICE, Customs and Border Patrol and other DHS agencies to identify and track immigrants. And DHS has pressured to do more in AWS. In a request for information from 201
7, DHS requested “Thinking, Modern Development, Security and Operations (DEVSECOPs) Information, especially in conjunction with Big Data, Analytics, PersonCentric, Entity Resolution and Machine Learning … to build, improve and support systems in big cloud environments, especially Amazon Web Services (AWS). ”
In a June 15 email to a DHS receiver, Amazon Web Services, Federal Sales Manager for Homeland Security, followed a conversation about Amazon’s Artificial Intelligence (AI) and Machine Learning (ML) which occurred at Redwood City, California, office of the consulting company McKinsey & Company. The company had a contract with “Integrated Consultancy Management Service” with ICE, a part that focused on “developing ICE [Enforcement and Removal Operations’] modernized vision and strategy.”
“We are ready and willing to support the important HSI mission” lead, whose name is edited in the edition, wrote. “I hope what we shared with … latest development updates about our AI / ML suite were of value.” The Amazon representative then noted “action articles” from the conversation, which meant that one could set up an “Innovation Workshop” focusing on a major HSI problem. “The list also included setting up a deeper technical briefing on a number of AI and ML technologies, including:
- Predictive analytical core competencies and installation scenarios.
- Elasticsearch as a managed service, [and] path to an ATO [Authority to Operate]
- Neptune Graph Database Value, User Cases and Security.
- Recognition Video Tagging / Analysis, Scalability, [and] Custom Object Library.
The action points mentioned an introduction to someone at USCIS to discuss the Agency’s Elasticsearch implementation
Ironically, the McKinsey encounter occurred when McKinsey employees spoke of the moral and ethical issues related to the company’s contract with ICE. A McKinsey contract with the ICE Department of Enforcement and Removal of Operations ended in July.
In July, the United States Civilian Freedom Union in northern California published results of a test they had carried out with help p of pictures of congressmen and a public database of 25,000 arrests of mughots. A total of 28 members of the Congress were flagged, including six members of Congressional Black Caucus, who match. The ACLU test followed a number of studies, including one of the US Government Accounts Account in March 2017, which found unacceptable incorrect pricing in face recognition systems. A MIT Media Lab study found that facial recognition systems often misidentified people with darker skin.
Amazon dispute the results of the ACLU test and tried to calm the DHS. In a follow-up message to the DHS contact, Amazon’s Sales Manager sent a link to an Amazon blog post stating “[Dr. Matt Wood’s] Comments on ACLU, Face Detection Accuracy and Confidence Levels” may be of interest in view of your ongoing efforts. “
In response to disclosure, the ACLU has repeated its call for “a moratorium on the use of face recognition technology for law enforcement and law enforcement purposes until the congress and the public debate on what is used by this technology should be allowed” and announced a request for information intelligence “which requires the Ministry of Home Affairs publishes about and how the Office uses or plans to use the technology. “
” ICE should not use face recognition for immigration or law enforcement, “said ACLU senior legislative adviser Neema Singh Guliani.” Congress has never allowed such use and should take immediate action to ensure that the federal authorities put the brakes on the use of recognition for immigration or law enforcement. “