Booz Allen Hamilton Holding Corp

Booz Allen Hamilton provides support to CBP’s National Targeting Center (NTC), a Virginia-based center that supports several DHS and law enforcement agencies by collecting and analyzing thousands of data entries used to target individuals. The center labels people as “high-risk” targets based on their history of leaving and entering the U.S. by cross-referencing multiple DHS-wide databases.  

CBP awarded Booz Allen Hamilton an $83 million, six-year contract in 2018 for providing the NTC with cyber threat software upgrades. 

In 2022, the NTC came under scrutiny over its role in spying on journalists, members of Congress, and other individuals not suspected of any wrongdoing. The operation, dubbed Whistle Pig, resulted in four congressional probes and an internal CBP review.

Booz Allen Hamilton developed a cloud-based analytics platform for ICE called the Repository for Analytics in a Virtualized Environment (RAVEn). RAVEn is a “big data” platform that performs analytics using artificial intelligence (AI) on large raw datasets so that ICE can make sense of this information in order to identify targets and predict patterns and connections between people and events. After awarding Booz Allen Hamilton a four-year, $31.2 million RAVEn contract in 2018, ICE awarded the company a RAVEn data analytics support contract in 2021. This contract is worth a potential $67.8 million and may be active until 2026.

RAVEn has access to several databases and links photographs, videos, documents, text, biometrics, and other types of data through machine learning models. RAVEn includes personal information such as location, financial data, telecommunications, biometrics, and biographic information on both U.S. citizens and immigrants. This information is obtained from U.S. government sources, such as law enforcement agencies, and commercial providers, such as license plate reader databases. RAVEn is hosted on Amazon Web Services, which is a key commercial partner of Booz Allen Hamilton.

Because RAVEn uses data that is already in possession of government agencies and private companies, rather than collecting new information, people are not notified when their information is stored in its database. RAVEn draws on information from other systems, such as the commercial phone surveillance platform Pen-Link, ICE’s Immigration Bond Management System, and the Peraton-designed Homeland Advanced Recognition Technology System (HART), DHS’s main biometrics database.

RAVEn also expands ICE’s facial recognition capabilities, as it serves as the portal through which ICE agents submit facial images to be recognized. These photos include mugshots, surveillance photos, and images confiscated from phones or other devices. RAVEn can also facilitate searches through the databases of other government agencies, like law enforcement, and tied to driver’s license records, reinforcing collaboration between police, Departments of Motor Vehicles (DMV), and immigration authorities.

The RAVEn platform poses serious privacy concerns beyond its facial recognition element. As with other DHS databases, information can be stored for up to 20 years after an investigation is closed. Bulk uploads of personal information, as is the case with most big data processes, is rarely closely supervised. Personal social media posts, mugshots, DMV information, geolocation data, and telecommunications records, among other personal details, can all be uploaded en masse, and ICE acknowledges that RAVEn may store inaccurate, out-of-date, or irrelevant information.

RAVEn replaces Palantir’s Investigative Case Management (ICM) system FALCON, which ingests massive amounts of data from ICE systems and provides visual and relational analysis that ICE agents use to target people and to plan raids. Despite Palantir’s claims that it does not assist ICE’s Enforcement and Removal Operations (ERO) with deportations, FALCON is a key tool used by ICE to power workplace raids, arrests, and deportations. Reporting has exposed the platform’s role in a number of high-profile raids.

Booz Allen Hamilton was also a contractor for ICM itself, under a 2014-2019 $21.9 million ICE contract. ICM was used, for example, in 2017 to track down the parents and other relatives of unaccompanied children crossing the U.S.–Mexico border, in what turned out to be a precursor to the Trump administration’s family separation policy.

Booz Allen Hamilton is a subcontractor on multiple ICE big data and surveillance projects. In 2018, the company was awarded a $143.6 million subcontract to work on the Homeland Advanced Recognition Technology (HART) system, a massive DHS database. Since 2021, the HART contract has been handled by Peraton.

In 2023, the U.S. Government Accountability Office published a report showing that HART was behind schedule, with information that DHS had failed to address “privacy weaknesses,” including risks associated with storing unique identifiable data such as DNA, fingerprints, facial patterns, full name, and country of origin. According to the report, HART will be fed with data from 140 partner agencies, including the Department of Justice, DOD, and Department of State, as well as federal, state and local law enforcement agencies.

HART was developed to replace DHS’s previous biometric database, the Automated Biometric Identity System (IDENT), managed by General Dynamics. According to DHS, HART will be hosted on Amazon Web Services, and all information currently held in IDENT will also be transferred to Amazon’s cloud services.

IDENT was created in 1996 and later became the key technology behind the Secure Communities program, enabling automated fingerprint sharing between local police agencies and ICE. Secure Communities was launched in 2008 and triggered a sharp increase in deportations during the early years of the Obama administration. By the time of the program’s suspension in November 2014, it was already responsible for an estimated 450,000 deportations.

Expanding on IDENT’s database of 230 million unique individuals, HART will become one of the largest biometric databases in the world, with the ability to store information on over 500 million individuals and support the collection of at least 720,000 data points daily. The types of biometric data HART collects and retains—for up to 75 years—include facial images, DNA, iris scans, fingerprints, and voiceprints. Additionally, DHS plans to have HART collect information about people’s “relationship patterns” in order to identify “political affiliation, religious activity, and familial and friendly relationships.” This information is accessible to CBP and ICE, as well as state and local police agencies.

CBP and ICE use mobile biometric devices to “identify faces and capture face data in the field,” allowing agencies to track people in public places without their knowledge or consent. This technology, in tandem with tracking political affiliation and religious activity via HART, might deter people from exercising their rights to political expression and assembly, according to the Electronic Frontier Foundation.

Another concern is that HART will disproportionately profile, surveil, and monitor people of color. FBI and MIT research has shown that current facial recognition systems misidentify people of color and women at higher rates than white people and men, and that misidentification increases for people with darker skin tones, leading critics like the ACLU to raise concerns that HART will encourage racial profiling. DHS’s tests of its own facial recognition system found high levels of inaccuracy, and a 2005 study found that “42 percent of [FBI] immigration hits in response to police queries were ‘false positives.’

Booz Allen Hamilton also worked to develop use cases, features, and requirements for ICE’s EDDIE biometric mobile application, which “collects and shares – ​in real time – fingerprints, facial scans, location information, and immigration history during immigration raids.” EDDIE feeds the acquired data into surveillance databases such as HART to check for matches. ICE raids result in many “collateral” arrests of non-targeted individuals, which means that anyone stopped by ICE in the course of a raid can be subject to biometric profiling.