Enemies Within: Inside the NYPD's Secret Spying Unit and bin Laden's Final Plot Against America

by Matt Apuzzo

  In New York, Cohen and Sanchez reviewed the dossiers that had been built on the 9/11 hijackers. Some of the intelligence came from government sources, but in the wake of the attacks, journalists from around the globe worked to piece together the lives of the terrorists aboard the airplanes that morning. Sanchez looked at the life of Mohamed Atta, the ringleader of the hijacking operation, and saw a learning opportunity for the NYPD. Here was a man who’d managed to fade into anonymity on three continents, someone who Osama bin Laden trusted to avoid detection while planning the most ambitious terrorist mission ever. He was a man who announced calmly, “Everything will be okay,” and then steered American Airlines Flight 11 into the North Tower of the World Trade Center. And he was about to become a case study in how to prevent terrorism. Sanchez and Cohen believed the Atta case contained the clues that future investigators could use to identify people before they attacked.

  • • •

  In Cairo, Egypt, Atta was raised in a family dedicated to scholarship, not prayer. His father, a lawyer, expected his son to learn. Atta and his sisters were not allowed to play outside. His parents timed the walk home from school and expected Atta to be back in the apartment studying without delay. In a densely populated, neighborly section of the city, his family seldom socialized or broke the fast with neighbors during the holy month of Ramadan.3 Eventually Atta graduated from Cairo University with a degree in architectural engineering and then continued his studies in Germany.

  Friends in Cairo don’t recall seeing Atta’s family at the mosque. They certainly weren’t regulars. But when Atta arrived in Hamburg in 1992, he immediately sought out the nearest one.4

  At the Technical University of Hamburg-Harburg, Atta applied himself to his studies. He also became increasingly religious and confrontational over moral and spiritual issues but never advocated violence.5 When he moved into university housing, his strict religious practices and stern personality quickly isolated him from his two successive roommates. He started a Muslim student group, a daily prayer session where investigators believe he met two men who would become hijackers with him.6 He began attending the Al Quds mosque, where a radical version of Sunni Islam was preached.

  In 1998 he rented an apartment with two friends, Ramzi bin al-Shibh and Marwan al-Shehhi. They shared a growing anti-American sentiment that, at least at first, they were not shy about discussing at the local pub where they went to talk—never to drink.7 The shopkeeper on the corner near Atta’s apartment recalls him growing out his beard and dressing in traditional Arab robes, called dishdashas.8 Al-Shehhi would go on to pilot United Airlines Flight 175 into the South Tower of the World Trade Center. Bin al-Shibh, who was unable to get a visa to join his friends on their mission, was later captured and imprisoned in secret CIA prisons. His cooperation provided the foundation for much of The 9/11 Commission Report.

  When Atta and al-Shehhi arrived in the United States in 2000, they stayed in New York, moving from cheap motels to short-term leases in Manhattan’s Hudson Heights and Brooklyn’s Park Slope neighborhoods.9 Al-Shehhi enrolled in an English class. They traveled the country, conducting premission surveillance, assembling their team, and training to fly jets. They attended mosques sporadically, visited internet cafés, and joined gyms.

  For Sanchez, this was a road map for the new NYPD. The federal government was tightening security at airports, getting tough on visa requirements, freezing money used to finance attacks, and requiring background checks for foreigners attending flight schools. With Cohen, the NYPD could go further. The story of Mohamed Atta was one of missed opportunities. There were people who’d seen signs of trouble, radical ideology, and anti-American vitriol: housemates and roommates, shopkeepers and pub patrons, fellow students and mosque-goers. They didn’t think anything of what they saw until it was too late.

  If the NYPD had its own eyes and ears in those cloistered communities, maybe things could be different. They needed be in the shop to spot the next Mohamed Atta in his kaftan with his newly grown beard. They had to be at the restaurant to overhear the group of friends ranting about America. If NYPD detectives infiltrated Muslim student groups, maybe they could identify the young man with the seething fanaticism. If the cops had a better handle on what went on inside the mosques, or which internet cafés were nearby—or even which gyms a young Middle Eastern man would attend—then maybe they could piece together the clues. Maybe they could prevent the next 9/11.

  The nearly successful effort by Richard Reid to detonate explosives in his shoes aboard American Airlines Flight 63 in December 2001 further validated the NYPD’s plans. Reid, who later became known as the “shoe bomber,” was a British citizen and Islamic convert who attended north London’s fiery, anti-American mosques. He had spent days in Paris before his flight, staying in the diverse neighborhood near the busy Gare du Nord train station. He ate at many of the restaurants in the area and used an internet café to send his mother his will and a final letter: “What I am doing is part of the ongoing war between Islam and disbelief (and as such a duty upon me as a Muslim).” Only after the attempted bombing did French authorities unravel the terrorist organization in Paris that supported Reid. To Cohen, that underscored the crisis.10

  Profiling is a loaded word in policing because it conjures images of white cops pulling over young black men and searching for guns or drugs. Racial profiling uses race as a stand-in for behavior: “That driver is probably up to no good because he’s black.” But racial profiling and behavioral profiling are different. The FBI, for instance, builds profiles of serial killers through its Behavioral Analysis Unit. And while such social science has not been immune from criticism, these profiles have been credited with helping solve crimes and catch killers.

  Sanchez envisioned a similar role for the NYPD, but with an important difference: It would not wait until a crime was committed. He wanted NYPD detectives to be the surrogates for all the people who missed the significance of Atta’s growing radicalization. It was an audacious plan, because the behaviors to be profiled were common not only to Atta and his murderous friends but also to a huge population of innocent people. Most café customers, gym members, college kids, and pub customers were not terrorists. Most devout Muslims weren’t, either. Nevertheless, Cohen liked the idea. He compared it to raking an extinguished fire pit. Most coals would be harmless and gray. But rake them carefully, and you might find a smoldering ember—a hot spot waiting to catch fire.11

  Like Sanchez’s very relationship with the NYPD, there was nothing like what he was proposing anywhere in American law enforcement. People who kill abortion doctors or bomb clinics have common behavioral traits too. They tend to be Christian; usually fervently so. They attend church, often participate in protests outside clinics, and acquire weapons. There has been no known effort to establish police eyes and ears in Christian churches, antiabortion groups, and gun clubs in hopes of spotting the next abortion-doctor killer. But New York was not under attack by fanatical, antiabortion Christians.

  There was, however, precedent for what Sanchez wanted to do. The surveillance abuses of the 1960s and 1970s were born out of a similar desire to identify trouble spots by monitoring lawful communities. Decades before that, in 1919, New York state senator Clayton Lusk led the Joint Committee Investigating Seditious Activities. He commissioned ethnic maps of New York. Irish, Germans, Russian Jews, Italians, and other groups were designated on color-coded charts to help authorities root out disloyalty and radicalism.12

  But Sanchez didn’t get his inspiration from New York’s troubled past. Rather, he got the idea from one of America’s closest allies, a country that had lived under the threat of terrorism for decades. Sanchez told friends and colleagues that the NYPD was taking its cue from Israeli officers’ methods of keeping tabs on the military-occupied West Bank, the swath captured from Jordan in the 1967 Six-Day War.13

  Sanchez’s proposal ignored some important differences between the US and Israel. Brooklyn and Queens, for instance, were not occupied t
erritories or disputed land. There was no security wall being erected in New York City. Israel does not have a constitution, and Muslims there do not enjoy the same freedom as Jews. In fact, they are routinely discriminated against.14 And, most significantly, unlike Israel, New York was not trying to preserve a religious identity.

  In the words of one former senior police official, reflecting on his role in transforming the NYPD, “Desperation breeds novel ideas.” Besides, it was hardly unusual for Israel to serve as a model for a US counterterrorism program. In the months after 9/11, American politicians flew to Israel in droves and extolled the virtues of Israeli tactics. Twenty years before the CIA opened its network of secret prisons, Israel was operating its own black site, called Facility 1391, to hold and interrogate prisoners indefinitely. Like its CIA cousins, Facility 1391 permitted harsh interrogation and was off-limits to human rights inspectors with the International Committee of the Red Cross.15

  The US looked to Israel, too, when crafting the rules for interrogation at black sites. In its memos, the Justice Department noted that the Israeli Supreme Court had, in 1999, determined that sleep deprivation, painful stress positions, and intense, lengthy interrogations were cruel and inhuman but did not constitute torture. The Justice Department concluded that international law allowed “an aggressive interpretation as to what amounts to torture.”16

  In fact, America’s signature offensive counterterrorism strategy not only replicated a tactic used by Israel but also used a strategy that the United States abhorred until 9/11. It was Israel that popularized the phrase targeted killing to describe its precise attacks on suspected militants. In July 2001 the American ambassador to Israel, Martin Indyk, had a different word for it. “The United States government is very clearly on the record as against targeted assassinations,” he said. “They are extrajudicial killings, and we do not support that.”17 That was before 9/11, before the Predator drone became the CIA’s signature weapon in the war on terrorism and before the word assassination was scrubbed from the US counterterrorism lexicon.

  Once Cohen persuaded Judge Haight to relax the Handschu rules, Sanchez’s vision could become a reality. The new rules made it explicit: “For the purpose of detecting or preventing terrorist activities, the NYPD is authorized to visit any place and attend any event that is open to the public on the same terms and conditions as members of the public generally.” The only caveat was that police couldn’t document and keep any information from these visits unless it related to potential criminal or terrorist activity. That rule was intended to prevent the NYPD from building files on innocent people, as it did during the 1960s. Cohen, however, took a very broad view of what qualified as information related to terrorist activity.

  Cohen and Sanchez enjoyed one advantage at the NYPD that they never had at the CIA. The department drew recruits from one of the most diverse talent pools in America, and the force reflected that. The FBI and CIA struggled to recruit Arabic speakers and Middle Eastern agents. In part, that was because those jobs required top-secret security clearances, which meant passing background checks that look unfavorably on applicants who still had strong ties overseas. The NYPD didn’t have that problem. The police force had long been a stepping-stone to the middle class for immigrants. One in five academy graduates were born overseas. So when Cohen went searching for officers who could blend into Muslim neighborhoods, he didn’t have to look far. He recruited young Middle Eastern officers who spoke Arabic, Bengali, Hindi, Punjabi, and Urdu. They would be the ones raking the coals, and inevitably they became known as “rakers.”18

  The effort began simply enough, with a copy of the 2000 US census. The police did what anyone else could do with that data trove. They mapped the city based on ethnicity and ancestry. The NYPD was interested in what it called “ancestries of interest.” There were twenty-eight, nearly all of them Muslim countries. There were Middle Eastern and South Asian countries such as Pakistan, Iran, Syria, and Egypt. Former Soviet states like Uzbekistan and Chechnya were included, too, because of their large Muslim populations. The last “ancestry” on the list was “American Black Muslim.”19

  Every day, the rakers would set out from the intelligence offices at the Brooklyn Army Terminal. They’d work in teams, usually of two, and visit businesses.

  They were not officially working undercover. At the NYPD, that designation is reserved for officers who use an assumed identity, with fake paperwork and a cover story. But the rakers weren’t advertising their police affiliation, either. Their job was to blend in and look like any other young men stepping in off the street.

  The routine was almost always the same, whether they were visiting a restaurant, deli, barbershop, or travel agency. The two rakers would enter and casually chat with the owner. The first order of business was to determine his ethnicity and that of the patrons. This would determine which file the business would go into. A report on Pakistani locations, for instance, or one on Moroccans. Next, they’d do what the NYPD called “gauging sentiment.” Were the patrons dressed in the clothing of observant Muslims? What were they talking about? If the Arabic news channel Al Jazeera was playing on the TV, the police would note it and also observe how people were acting. Were they laughing, smiling, or cheering at reports of US military casualties in Iraq and Afghanistan? Did they talk Middle Eastern politics? If the business sold extremist literature or CDs, the officers would buy one or two. Was the owner selling fake IDs or untaxed cigarettes? Police would note it. If customers could rent time on a computer, police might pay for a session and look at its search history. Were people viewing jihadist videos or searching for bomb-making instructions?

  On their way out, the rakers would look for bulletin boards or fliers about community events. Was there a rally planned in the neighborhood? The rakers might attend. Was there a cricket league? The rakers might join. If someone advertised a room for rent, the cops would bring the flier back with them. That could be the cheap apartment used by the next Mohamed Atta.

  • • •

  In the beginning, raking was normally done by neighborhood. Sanchez had the NYPD carve the city up into about eighteen zones, and the rakers would visit Muslim businesses in each. They often picked their own targets, with a supervisor sitting in a parked car somewhere nearby in case of trouble. Sometimes they were sent to neighborhoods based on world events. If there was a car bombing in Lebanon, a Predator drone strike in Pakistan, or a firefight in Afghanistan, the rakers would be in those neighborhoods, gauging sentiment and reporting back. If people in a Pakistani barbershop were enraged over a drone attack that killed nearby civilians, it might be a warning sign that retaliation was imminent.

  The rakers were in mosques too, gauging the sentiment of the imams and the congregations. They’d scan bulletin boards for scraps. They bought neighborhood newspapers and identified religious schools, community centers, hotels, and gyms. The NYPD was creating a new kind of map. Just as political maps showed the borders of New York City and topographic maps revealed the city’s elevation, rakers charted New York’s human terrain, mapping people and their attitudes.20

  The idea of getting to know a community was a hot topic at the FBI too. And like the NYPD, the bureau had its own advisor on loan from the CIA, an analyst named Phil Mudd. He wanted the FBI to be more aggressive, to focus less on making isolated criminal cases and more on collecting intelligence. Inside the bureau, he was one of the biggest advocates of what became known as domain management, a process in which FBI offices nationwide compiled information on communities and assessed where terrorist threats might emerge. Like the NYPD, the FBI began with census data. It could then overlay other data—crimes, informant locations, potential targets—and create maps of neighborhoods. It used that information to find informants, assess threats, and decide where to conduct outreach to community leaders. Also like the NYPD, the FBI focused its efforts on Muslim neighborhoods. There was no FBI Catholic outreach program, nor was anyone interested in mapping Scottish immigrant neighborhoods. Domai
n management was controversial both inside the FBI and, when it became public, among civil libertarians.

  While the NYPD and FBI had similar goals, they diverged at one important point. The FBI was prohibited, both under its guidelines and under federal law, from collecting and storing information concerning constitutionally protected activities such as religious and political speech unless related directly to law enforcement activities. That meant the FBI could not keep tabs on which pastry shop posted religious fliers on its bulletin board. Nor could the bureau put in its files that a Turkish couple owned a restaurant that served a “devout clientele.” The FBI could not keep a file on an Egyptian travel agent who was “devout in appearance.” And it could not send plainclothes agents into mosques to assess “sentiment.”

  NYPD did all of that. While Cohen had promised Judge Haight that his reinvented Intelligence Division would follow the same rules as the FBI, it did not. The newly created NYPD unit was explicitly instructed to “analyze religious institutions, locations, and congregations.”21

  Sanchez took a particular interest in the program’s success, reading the reports and coaching the police on how to improve. Their daily dispatches were compiled into bound color reports that filled the bookshelves in Cohen’s office. Mosques and religious schools were catalogued and hot spots were mapped by ethnicity for every precinct. The reports allowed police to visualize their city in a new way. If a group of young Muslim men were growing increasingly radicalized and planning an attack, these hot spots were the likeliest places to detect and locate them. There were hundreds of hot spots on the maps, all screaming for attention.