Among the signals that should have alerted the FBI well before September 11, 2001, that Islamic terrorists might be thinking of crashing airplanes into American buildings was the 1996 confession of a captured Pakistani terrorist named Abdul Hakim Murad. He and others had planned to blow up 12 U.S.-owned airliners over the Pacific Ocean—and he had taken flying lessons in the U.S. to prepare to crash a plane into CIA headquarters. The planned suicide flight was not included in the criminal charges against Murad, apparently because it had not ripened into a provable conspiracy. But surely a crack counter-terrorism agency would have gone on the lookout for any signs of similar plots by other jihadists.
What did the FBI do? It "effectively forgot all about it," asserted Sen. Richard C. Shelby of Alabama, who was the Senate Select Intelligence Committee's senior Republican for the past six years, in an impressive 84-page brief last month calling for "radical reform" of the intelligence community. "Convinced that the only information that really matters was information directly related to the criminal investigation at hand, the FBI thus ignored this early warning sign that terrorists had begun planning to crash aircraft into symbols of U.S. power. Thus, rather than being stored [and] assessed and re-assessed in light of a much broader universe of information about terrorist plans and intentions over time, the Murad data-point.... slipped out of the FBI's usable institutional memory." So it was that in the summer of 2001, it never crossed anyone's mind at the FBI to see the accumulating evidence of possible Qaeda plans to crash planes into buildings as part of a pattern dating back for years.
Shelby's point was not merely to detail yet another example of the appalling succession of FBI intelligence failures in the decades before the 9/11 attacks. His broader thesis, also suggested by other influential lawmakers and experts in recent months, was that "the bureau—as a law enforcement organization—is fundamentally incapable, in its present form, of providing Americans with the security they require against foreign terrorist and intelligence threats."
From this premise, Shelby and others reason that perhaps we should strip the FBI of its counter-terrorism and counter-intelligence functions, and move them to a new domestic intelligence agency. It would have no guns, no badges, no arrest authority, no law enforcement responsibilities—and none of the incentives that have long pervaded the law-enforcement community to let prosecutorial priorities eclipse the need to collect and analyze as much information as possible about possible terrorists before they can strike.
Advocates of radical change include an independent advisory commission on terrorism chaired by former Virginia Gov. James S. Gilmore III, a Republican, and Sen. John Edwards, D-N.C., who may push the idea in his nascent presidential campaign. Sens. Bob Graham, D-Fla., the former Intelligence Committee chairman, Charles E. Schumer, D-N.Y., and other lawmakers have also expressed doubts that the FBI is up to the job.
The model some critics tout is Great Britain's domestic intelligence agency, MI5. It is separate from (although it cooperates with) both Scotland Yard, which like the FBI is the top criminal investigative agency, and MI6, which like the CIA collects intelligence overseas. The domestic intelligence organizations in Canada, France, Germany, and Israel are also separated from their law enforcement agencies; the U.S. is unusual in combining these functions in one agency.
Shelby's 84-page analysis details his view that cops cannot do the work of spies: "Intelligence analysts tend to reach conclusions based upon disparate fragments of data derived from widely-distributed sources and assembled into a probabilistic 'mosaic' of information. They seek to distinguish useful 'signals' from a bewildering universe of background 'noise' and make determinations upon the basis of vague pattern recognition, ... context, and history.... Intelligence analysts think in degrees of possibility and probability, as opposed to categories of admissibility and degrees of contribution to the ultimate criminal-investigative end of proof 'beyond a reasonable doubt.'"
Law enforcement agents, on the other hand, are trained to pursue "a discrete bundle of information ... to prove elements of crimes against specific potential defendants in a court of law." This, says Shelby, helps explain the FBI's history of "positive aversion to long-term strategic analysis of the sort routinely expected of intelligence agencies" and "its inability to assess what is in its own files."
Shelby's case for taking domestic spying out of the FBI seems powerful. Alas, some of the objections do too, as I have learned while researching a column initially conceived as a pitch for an American MI5.
The least compelling objection may be the most common one. It is that a new domestic intelligence agency, untethered from the need to focus on prosecutable crimes, would run roughshod over civil liberties. To the contrary, a new agency might be good for civil liberties. Explains Robert Litt, a senior Justice Department official under the Clinton administration who is "agnostic" about the MI5 model: "To meet the terrorist threat today, investigators are going to collect and analyze more information than ever before. The opportunities for misuse of that data may be more limited if it is gathered by an organization that is not responsible more generally for enforcing the criminal law." If we don't want a secret police, maybe we should put the secrets and the police in different agencies.
The more compelling objection to an American MI5 is that it might prove less effective in preventing terrorist attacks than a reformed FBI—"a step backward in the war on terror, not a step forward," in the words of FBI Director Robert S. Mueller III. Some independent experts agree, including John L. Martin, who headed the Justice Department's internal security unit before retiring in 1997. A new agency "merely creates a new piece of turf," says Martin, adding that MI5 has had problems akin to the FBI's.
A new agency would also, Mueller argued in a December 19 speech, resurrect the recently demolished "wall" between law enforcement and intelligence, which had contributed to the government's failure to prevent the 9/11 attacks by sharply limiting information-sharing and cooperation both within the FBI and between the FBI and CIA. Now, Mueller said, the FBI can "make strategic and tactical choices between our law enforcement options of arrest and incarceration and our intelligence options of surveillance and source development," without turf wars between cops and spies.
The most daunting problem with any plan to take domestic spying out of the FBI may be the difficulty of changing horses in midstream, amid a torrent of jihadist terrorism. It would take years of disruptive bureaucratic reconfiguration to replicate the FBI's enormous investigative infrastructure around the country and the world—more years, perhaps, than it would take to change its deeply ingrained cops-and-robbers culture. And Mueller says that that effort is well under way.
Is it? More than 13 months after 9/11, Mueller's second in command, Bruce J. Gebhardt, sent the FBI's 56 field offices an e-mail (later leaked to reporters) complaining that he was "amazed and astounded and at a loss to understand" why some offices were still unresponsive to Mueller's orders that they treat counter-terrorism as the overriding priority. Agents "need to get out on the street and develop sources," with a greater "sense of urgency," he wrote. Mueller himself found it necessary to warn in a November 29 memo (also leaked) that he would not tolerate "bureaucratic intransigence."
From a political standpoint, radical reform seems a long shot for the time being. Homeland Security Director Tom Ridge suggested last November—after meeting in London with the head of MI5—that the MI5 model was "different than anything we would want to have done in this country." Whether his words reflect an administration consensus is unclear.
Meanwhile, Shelby has a fallback proposal: To separate the FBI's domestic intelligence people "into a semi-autonomous organization that reports to the FBI director for purposes of overall coordination and accountability, but which would in every other respect be wholly separate from the 'criminal' components of the FBI."
Is that a good idea? Is an American MI5? Good enough, in my view, to warrant careful congressional hearings on the whole range of organizational options and open-minded consideration by the administration. Policy makers, no less than agents, should take to heart something that Gebhardt said in his e-mail to FBI field offices: "You cannot fail at this mission. Too many people are depending on us."