President Trump has inherited a vast domestic intelligence agency with extraordinary secret powers. A cache of documents offers a rare window into the FBI’s quiet expansion since 9/11.
At its lowest level of investigative activity, on the basis of vague tips or broad intelligence interests, the FBI can follow people with airplanes, examine travel records, and analyze links between email, phone, and other records collected by intelligence agencies.
Two large FBI manuals obtained by The Intercept, one of which is classified, offer previously unreleased information about just how powerful an intelligence apparatus the FBI draws on even for low-level checks, known as “assessments.”
Assessments allow agents to look into tips or leads that don’t meet the standard for opening an investigation, which requires specific information or allegations of wrongdoing — an “articulable factual basis” for suspicion, as FBI rules put it. In an assessment, by contrast, an agent just needs to give an “authorized purpose” for their actions. Agents can open assessments “proactively,” in order to evaluate potential informants, collect intelligence about threats surrounding public events, study a field office’s geographical area, or gather information about a general phenomenon of interest to the bureau.
When assessments were made an official category in the last months of the George W. Bush administration, civil liberties advocates warned that the change eroded limits imposed on the bureau after the abuses of J. Edgar Hoover’s FBI came to light in congressional investigations in the 1970s. Those limits set high standards for the FBI’s investigations, requiring them to be tied to evidence of wrongdoing and adding layers of oversight. Advocates now say that the standard for opening an assessment is just too low, allowing for enormous amounts of information to be collected and retained even when nothing turns up.
The FBI counters that assessments offer a common-sense, expeditious means of handling tips and doing due diligence on potential threats. The bureau maintains that assessments are still subject to oversight and are usually quickly closed if they don’t meet the threshold for a more intensive investigation.
The Intercept is publishing in full for the very first time the Domestic Investigations and Operations Guide, or DIOG, which governs FBI operations. The DIOG’s rules for assessments, which were previously released with significant redactions, make clear that the FBI can follow up on a tip or complaint by probing government records and conducting interviews. The bureau need not necessarily keep a record of this search. But if agents decide to dig deeper by opening an assessment, they are allowed to have informants collect information, and they can also physically surveil the subject — including by airplane. In some cases, they can issue grand jury subpoenas. If the purpose of an assessment is to evaluate someone as a potential informant, agents can give polygraph tests, dig through trash, and use fake identities in the course of their research.
When an assessment is linked to counterterrorism, agents are required to address a sort of checklist of questions called the “baseline collection plan.” The American Civil Liberties Union obtained the 2009 baseline collection questions through a public records request several years ago. The Intercept has now obtained an unredacted version of the FBI’s Counterterrorism Policy Guide, which reveals new details about baseline collection and the resources the FBI uses to carry it out.
According to this guide, as part of baseline collection, the FBI avails itself of an array of sophisticated databases and analytical tools. These include “Clearwater” and “Polaris,” tools that are used to analyze links between email, phone, and other records collected by government agencies through various programs, and about which little has been publicly acknowledged. The policy guide directs agents to query the subject’s telephone numbers in those and other systems, including the FBI’s Telephone Application, used to analyze phone records; the Investigative Data Warehouse, which as of 2008 contained some 1 billion unique records and has been described as the FBI’s Google; the Data Warehouse System, which collates many FBI investigative records; and the Data Loading and Analysis System, which holds files from digital media seized by the FBI, including cellphones, computers, CDs, and the like.
The FBI also traces the subject’s travel history through Department of State visa and passport records, a Customs and Border Protection database, and data held by a private company called the Airlines Reporting Corporation, which manages itineraries for airlines and travel agencies. In 2009, Wired reported that the FBI was seeking access to ARC’s full database, which would include billions of travel records showing the data printed on the front of an airplane ticket and the method of payment used. ARC told Wired at the time that it only released records in response to a legal order such as a national security letter or subpoena, and that such demands had so far resulted in 17,000 records being turned over. It is unclear from the policy guide what sort of access the FBI now has to ARC, or whether it is checking previously obtained documents. ARC declined to comment.
Agents running assessments can further check the databases of the Financial Crimes Enforcement Network in the Department of the Treasury. They can ask for information on gun purchases from the relevant local bureau of Alcohol, Tobacco, Firearms and Explosives. They can run subjects’ names through the National Counterterrorism Center, and even through the CIA and NSA.
An unclassified October 2013 communiqué from the FBI’s counterterrorism division also suggests searching a classified terror watchlist, an FBI internet data-mining tool called Base Jumper, DMV files of driver’s license photos, and state-level business filings. The communiqué also lists open source resources, including social media sites, “web crawlers/deep web engines/social media searchers,” and photo websites like Flickr, Picasa, and Shutterfly.
The document adds that “in addition to queries of databases/resources and interviews, careful consideration should also be given to utilizing other investigative methods authorized in Assessments, such as [confidential human source] tasking and physical surveillance, when appropriate.”
The FBI, through a spokesperson, said that assessments were limited in scope and subject to rigorous oversight: “Regular file reviews, which include a review of whether or not the assessment should be closed or should continue for another 60- or 90-day period, as well as realistic resource constraints provide the necessary safeguards to ensure that individuals are not subject to long-term surveillance,” said an FBI spokesperson.
Yet, critics note that it appears the data collected can be retained for decades and queried for years, even when it does not lead to an investigation. The FBI would not clarify the precise limits on how long it can keep assessment data.
“Everyone should know that the FBI claims the authority to use extremely intrusive methods to investigate people without any factual basis to suspect that they have engaged in actual wrongdoing,” said Hina Shamsi, director of the National Security Project at the ACLU. Those methods “can reveal intimate information about how people live their lives,” she said, adding, “This is an ineffective authority because it clogs up the system with baseless investigations, and it is a dangerous authority because it has been abused to spy on innocent people exercising their First Amendment rights to protest, worship, and engage in political advocacy.”
Despite initial condemnation from civil liberties groups, in the years since assessments were instituted, they have received relatively little attention — except when the FBI is criticized for not having caught in an assessment someone who went on to carry out an attack.
The DIOG specifies that an assessment can’t be opened on the basis of “arbitrary or groundless speculation,” and the agent is supposed to be able to articulate why the assessment isn’t “frivolous or improper.” Assessments are also not supposed to be based “solely” on First Amendment-protected activities, or on the target’s race, religion, or ethnicity — although such characteristics can be taken into account. The Counterterrorism Policy Guide indicates that an assessment isn’t enough to place someone on the terror watchlist.
The examples of assessments given in the DIOG — blacked out in the versions that have been made public previously — show how FBI agents are expected to understand these boundaries.
In one example, if an FBI agent gets a tip that men of Middle Eastern descent rented a boat and asked the marina owner to circle military installations and nuclear power plants on a map, he can open an assessment and look into the men’s backgrounds. The example emphasizes that “the men’s national origin is not the motivating factor for the assessment” — unless other information suggests it is relevant, “for example, if existing intelligence reporting suggested that Al Qaeda was planning an attack using boats.” In another example, the FBI describes an assessment asking whether a public event — such as an auto show — could draw national security threats.
According to the DIOG, some assessments can take whole neighborhoods into their sights, with agents collecting information on the “composition of the community, the different ethnic groups, religious affiliations, community interests and dynamics, businesses, etc. for analysis and planning.” This kind of assessment could involve checking public, commercial, and governmental data, including visa information, and the FBI field office could “store, analyze, map, and share the information.” Such practices have stoked controversy, particularly among civil liberties and Muslim groups, who protest that such assessments risk tarring whole communities with criminal associations.
According to the last available figures, obtained by the New York Times, the FBI opened more than 82,000 assessments between 2009 and 2011. Most of them were closed without developing into investigations. Critics seized on these numbers as evidence that the FBI was casting too wide a net. But the FBI protested that the opposite was true: Assessments allowed agents to be more discriminating, the FBI maintained, and to open fewer preliminary investigations, which are even more invasive.
Between the 2008 and 2011 versions of the DIOG, the rules on assessments actually got more permissive, granting agents more flexibility in using surveillance teams, for example, and allowing them to run checks in commercial and law enforcement databases without formally opening assessments.
“There’s been a lot of morphing on assessments,” Jeff Danik, a former FBI agent who worked on counterterrorism, told The Intercept. “At first it was very cut and dry when it says, ‘We can’t do this.’ Now it’s more, ‘We’re not going to go against what it says, we’re just going to get the approvals, which are going to be given pretty freely.’ It’s kind of found its own middle ground. They didn’t take the rules away, but they made it a little easier to go forward.”
There is no time limit on how long an assessment can last, though the FBI stressed that supervisors must periodically review the basis for keeping it open. Danik said there was often pressure to either “convert it to an investigation or get it closed.”
Assessments have mostly come to the public’s attention after domestic terror attacks, when the bureau gets criticized for not having caught perpetrators beforehand. After last year’s bombings in New Jersey and New York, for example, the public learned that the suspect had been scrutinized in an assessment. Yet nothing had turned up that warned of his future violence or warranted either an investigation or an arrest.
According to some critics, assessments offer a worst-of-both-worlds solution, in which agents pursue every lead, generating overwhelming caseloads, even while the bureau is unfairly blamed for not having followed up beyond what it was legally allowed to do.
Editor-in-Chief: Betsy Reed. Series Editors: Laura Secor, Ryan Tate. Associate Editor: Andrea Jones. Reporters: Trevor Aaronson, Cora Currier, Jenna McLaughlin, Alice Speri. Research: Alleen Brown, Talya Cooper, Danielle Mackey, Eseosa Olumhense, Miriam Pensack, John Thomason. Art Direction: Stephane Elbaz, Philipp Hubert, Nick Simmons. Additional Photo Editing: Soohee Cho, Shaun Lucas, Chelsea Matiash. Development: Tom Conroy, Andy Gillette, Carl Licata, Cacie Prins, Raby Yuson.