This Secret $35 Million FBI Unit Mixes Facial Recognition With Big Data To Investigate America’s Most Horrific Crimes
Big Data – In the wake of a mass shooting or terror attacks investigators can be left with hours of CCTV footage, video from witnesses, or clips from social media. Take the Boston Marathon bombing in 2013, where the FBI received over 13,000 videos and had 120 analysts probe them all for clues.
As it seeks to improve its ability to sift through such abundances of video at major crime scenes, Forbes has learned that a previously-unreported forensics division within the FBI called the Multimedia Exploitation Unit (MXU) has been tasked with this role. It has cost at least $35 million since 2016 and draws on cutting-edge expertise from Mitre Corporation, a non-profit government skunkworks that receives between $1 and $2 billion a year from the U.S. government.
Documents obtained via FOIA reveal that MXU, run out of the FBI’s Criminal Justice Information Services (CJIS) division, seeks to “process and exploit multimedia assets” so that the FBI can “transform… bulk data into investigative leads.” The unit has pushed one smaller security contractor – West Virginia-based Azimuth Inc – to create tech for “bulk multimedia search” and “image clustering” that brings together content that’s been flagged by facial and object recognition technology in an easy-to-digest tool.
There’s little more information available on MXU, though one conference line-up online revealed that the chief ,as of late 2018, was Sam Cava, previously head of the Department of Defense’s Biometrics Fusion Center, established in 2000 to help the military deploy technologies like facial recognition. Used to keep tabs on military detainees, prisoners of war and other individuals deemed a threat to national security, the center started collaborating with the FBI CJIS unit back in the mid-2000s.
Looking through contract records, Mitre was the biggest tech and services provider for the unit, receiving nearly $20 million in contracts. The only other major supplier was Azimuth Inc, from West Virginia, which received $15 million. The first contracts for the unit were handed out in 2016, according to a review of federal procurement records.
Chris Piehota, the recently-retired chief of the FBI science and tech division, tells Forbes that the unit is a small, specialist division within the FBI whose sole focus is to build software to analyze photo and video in major crimes.
“If you’re collecting terabytes of video information from a horrific scene or a horrific event, the MXU comes in with their tools. They help investigative personnel make that footage usable. They parse out what we call non-pertinent footage. And they provide analytics and identification to pertinent investigative information in that video footage,” Piehota adds. “They were key in all of the mass shootings that happened over the past several years in the United States.”
The FBI acknowledged a request for comment on the MXU provided comment at the time of publication. Mitre had previously repeatedly declined Forbes’ requests for comment on any aspect of its work.
How American intelligence leans on Mitre
MXU is one of many collaborations between Mitre and the FBI. They’ve had a working relationship since at least the 1990s, when the elite government labs provider started to expand beyond its typical aviation and military work. A 1993 Boston Globe article detailed Mitre’s work to improve the National Crime Information Center, a huge electronic clearinghouse of crime data that’s accessible to most policing agencies. One addition was a system to send mugshots directly to patrol cars. “The FBI used Mitre extensively for all sorts of different things from development of processes, development of software, and then providing personnel to do various tasks like intelligence, cybersecurity, that sort of thing,” recalls Jason Truppi, a former FBI agent turned cybersecurity entrepreneur, who was at the agency between 2008 and 2015.
Earlier this week, Forbes revealed work on a Mitre-FBI project to take people’s fingerprints from images on social media sites like Facebook, Instagram and Twitter. The two also worked together to build the Next Generation Identification (NGI) system, said by the FBI to be “the world’s largest and most efficient electronic repository of biometric and criminal history information.” As described by Piehota, it’s a place where any law enforcement body in the U.S. can look up fingerprints, faces or other biometric information of criminal suspects.
Mitre engineers have also worked with the FBI to take down the Silk Road drugs marketplace, and other dark web investigations. As revealed in another document obtained via FOIA, Mitre has a close relationship with the FBI’s intelligence division. Summing up the importance of Mitre to the FBI is a contract with the agency’s Directorate of Intelligence, which reads: “Having had a close support relationship with the FBI for over 30 years, and in particular with the intelligence programs of the bureau for over 10 years, Mitre’s National Security Engineering Center provides a unique perspective and a demonstrated propensity to help solve complex challenges.”
Mitre’s work with the law enforcement goes beyond snooping, however. One additional contract obtained by Forbes details its work on protecting the FBI’s own data. It asks Mitre to ensure the FBI’s Information and Technology Branch to develop access rules for the agency’s “top secret, secret and unclassified” information troves. As much of Mitre’s own work for the FBI covers both targeted and mass surveillance, the nonprofit will be helping keep its own work under lock and key.
This article originally appeared on fool.com To read the full article and see the images, click here.
Nastel Technologies uses machine learning to detect anomalies, behavior and sentiment, accelerate decisions, satisfy customers, innovate continuously. To answer business-centric questions and provide actionable guidance for decision-makers, Nastel’s AutoPilot® for Analytics fuses:
- zAdvanced predictive anomaly detection, Bayesian Classification and other machine learning algorithms
- Raw information handling and analytics speed
- End-to-end business transaction tracking that spans technologies, tiers, and organizations
- Intuitive, easy-to-use data visualizations and dashboards