Summary  Why Section 206 Should be Modified  Terrorism is Not Just a Crime  Response  The Last Word?

Section 206.
Roving Surveillance Authority under FISA

A Summary by Mary DeRosa

Section 206 expands the authority of the Foreign Intelligence Surveillance Court (FISC) by allowing it to order "roving" or multi-point surveillance. Previously, FISA required a separate FISC authorization to tap each device a target used. Technically, the Court's order would direct a specific carrier or individual to assist in the surveillance. As a practical matter this required separate application to the FISC each time a target switched from pay phone to cell phone, email to Blackberry, etcetera. Now, the Court may order surveillance focused on the target, rather than the device he or she is using when "the actions of the target of the application may have the effect of thwarting the identification" of a specific device. This new order does not specify the person who is directed to assist; the government may serve the generic order on individuals or carriers that it identifies.

Section 206 modernizes FISA wiretap authority and brings it closer in line with the criminal wiretap laws, which have permitted roving wiretaps since 1986. The FISA roving authority differs from the criminal wiretap laws, however, in that it does not contain a requirement that the government "ascertain" where the targeted communication will take place before intercepting communications. The criminal wiretap law, 18 U.S.C. section 2518(12), states: "An interception of a communication under [a roving surveillance] order . . . shall not begin until the place where the communication is to be intercepted is ascertained by the person implementing the interception order."

Also, under FISA, the government need not always identify a target to obtain a warrant. FISA section 105(c)(1)(A), requires that an order specify "the identity, if known, or a description of the target of the electronic surveillance." Therefore, with the new roving surveillance authority, the government may obtain an order to conduct surveillance that specifies neither a named target nor a specific device to tap, although the order must provide a description of the target if a name is not known.

Why Section 206 Should be Modified
by James X. Dempsey

Section 206 of the USA PATRIOT Act authorized "roving' wiretaps under the Foreign Intelligence Surveillance Act (FISA). Under this provision, the FBI in an intelligence investigation can obtain a court order to intercept an individual's communications without specifying the particular phone line or computer account to which the interception will apply. Thus, if a suspected spy or terrorist is moving from safehouse to safehouse, or is using public access computers, or is changing cell phones frequently, the FBI can continue to intercept the individual's communications without having to return to court for a new order covering each new phone line or computer.

It makes perfect sense that the FBI should have roving tap authority in intelligence investigations of terrorists. In fact, the FBI had been given roving tap authority in criminal investigations of terrorists in 1986, and there was no good reason why similar authority had not been granted at the same time for intelligence investigations under FISA. During the PATRIOT Act debate, the Justice Department sometimes justified Section 206 by claiming that it needed the same authority to investigate dangerous terrorists that it already had for ordinary criminals, but in reality DOJ already did have -- prior to the PATRIOT Act -- roving tap authority for terrorists plotting attacks or raising money and being investigated under the criminal law.

So, as with so many provisions of the PATRIOT Act, the concern with Section 206 is not with the authority itself. Rather, the issue is the lack of adequate checks and balances. Roving taps pose obvious risks. Constitutionally, they are suspect, since they depart from the Fourth Amendment's explicit requirement that all warrants must "particularly describ[e] the place to be searched.' Practically, they pose the risk of intercepting the communications of innocent persons, risks that are separate from those addressed by the minimization requirement of the wiretap laws. Congress recognized when it first created the roving tap authority for criminal cases that it needed to add extra procedural protections to overcome the constitutional problems and to guard against the interception of the conversations of innocent people.

However, even though the DOJ argued in the PATRIOT Act debates that it wanted in intelligence investigations of terrorists the same roving tap authority it had for criminal investigations of terrorists, Section 206 lacks two important protections that apply to criminal roving taps. The first is the so-called ascertainment requirement—the requirement that agents actually ascertain the location of the suspect before turning on their recording devices. Secondly, some additional changes to FISA adopted outside of the normal process in the Intelligence Authorization Act a few months after the PATRIOT Act had the probably unintended effect of seeming to authorize "John Doe' roving taps—that is, FISA orders that identify neither the target nor the location of the interception.

As a result, a new imbalance has been created, leaving the FBI without appropriate oversight in the use of its roving tap power under FISA and making it more likely that the email and telephone conversations of innocent people will be inadvertently monitored. Fortunately, these are problems that can be fixed without taking away the roving tap authority for intelligence investigations.

First, roving taps in criminal cases are subject to an "ascertainment' requirement. This means that the surveillance can be set up in advance but cannot be activated until the officers in charge of the operation have a good sense that the target is actually using a particular telephone or computer. For example, the roving tap on a hotel room cannot be activated until the suspect checks into the hotel. A roving tap on a computer in a library can be all set up but it cannot be activated until the suspect is seen to sit down at the particular computer. This ascertainment requirement helps ensure a wiretap is turned on only when the target is present, and thereby protects the privacy of innocent users. The FISA roving tap authority added by the PATRIOT Act has no parallel protection.

This can be fixed pretty easily: an ascertainment requirement should be added to the FISA roving tap authority in 50 U.S.C. § 1805(c) to require that "in cases where the facility or place at which the surveillance is to be directed is not known at the time the order is issued, the surveillance be conducted only when the presence of the target at a particular facility or place has been ascertained by the person conducting the surveillance.'

This language was proposed in the SAFE Act, a bipartisan bill introduced in the 108th Congress as H.R. 3352 and S. 1709 and expected to be reintroduced in the 109th Congress. It is based directly on the ascertainment requirement for roving bugs in criminal cases, which requires that the interception "shall not begin until the place where the communication is to be intercepted is ascertained by the person implementing the interception order.' 18 U.S.C. § 2518(12). That standard is clearer than the ascertainment requirement currently provided for roving phone taps in criminal cases, which was rewritten a number of years ago without much forethought and no longer makes much sense.

Second, as to the John Doe issue, the rule for FISA taps should be the same as the rule for criminal taps, which is very simple: the government's application and the judicial warrant should specify either the person or the place to be surveilled. See 18 U.S.C. § 2518(11)(b)(ii), in the criminal wiretap statute, which specifically requires that a roving tap must "identif[y] the person' to be tapped.' However, as a result of changes made "outside the scope' of the FY 2002 intelligence authorization bill conference, FISA as now drafted seems to allow applications and orders that specify neither the person nor the location to be tapped. Specifically, the statute permits the FISA court to issue an order that specifies "the identity, if known, or a description of the target of the electronic surveillance' and "the nature and location of each of the facilities or places at which the electronic surveillance will be directed, if known.' 50 U.S.C. § 1805(c)(1)(A), (B). The roving tap authority, which was added in a somewhat careless fashion, does not require roving taps to specifically identify the target of the surveillance. Thus, if neither the identity nor the facilities are known, the order could be issued with only a general description of the individual to be tapped. Neither the statute nor the legislative history indicates how specific any such description must be.

This is unprecedented, probably unintended, and probably unconstitutional. It permits the FISA court to issue a wiretap order authorizing the FBI to listen in on a telephone or e-mail conversation of a person not named in the order, at any telephone or computer the unnamed person might use. This loophole, even if it was unintentional should be corrected. A simple change that clarifies that a FISA wiretap order must specify either the target or the facility—but not neither—would remedy this problem and should not be controversial. Again, the bi-partisan SAFE Act addressed this issue, adding language to 50 U.S.C. 1805(c)(1) making it clear that, if the identity of the target is unknown, the facilities and places shall be specified, or, if the facilities or places where the surveillance is to be conducted are unknown the identity of the target shall be specified.

The Justice Department has already admitted that a roving tap order must include a "description' of the target. See Letter from John Ashcroft to Sen. Orrin Hatch, at 2-3 (Jan. 28, 2004). But "description' is not defined in FISA and DOJ has failed to offer an adequate definition of its own. Is "20- to 35-year-old Arab male' an adequate description? It describes thousands of people in the United States. On the other hand, if a description is an alias or is so detailed that it could reasonably describe only one person, then that description surely could be considered an "identity,' permitting a roving tap to be employed.

This problem should be corrected legislatively. Clear legislative definitions are particularly important in FISA, because the role of the courts in defining the limits of the statute is much more constrained. Title III, in contrast to FISA, requires notice after the fact to all targets of surveillance. This means that persons mistakenly surveilled can meaningfully challenge the government's actions. Under FISA, the government normally never notifies targets that they have been wiretapped. Even in cases where FISA-gathered evidence is used in a criminal proceeding, the normal adversarial rules have not been applied: No defendant has ever gotten access to the underlying FISA application and order, thereby severely hampering their ability to challenge the legality of the search.

Like so many of the controversial surveillance provisions of the PATRIOT Act, Section 206 took an otherwise valid idea a step too far. Extending roving taps to intelligence investigations was not in itself a substantial civil liberties problem. But authorizing roving intelligence taps without sufficient safeguards is.

Terrorism is not just a crime
by Paul Rosenzweig

Shortly after the Patriot Act was passed, opponents of the Act wrote of section 206: "These wiretaps pose a greater challenge to privacy because they are authorized secretly without a showing of probable cause of crime... This Section represents a broad expansion of power without building in a necessary privacy protection.' Thus was painted an apocalyptic picture of Big Brother on steroids—a security apparatus that can listen to "anyone at anytime.'

Fortunately, the discussion of section 206 that we are engaged in here is a far more measured and thoughtful one of nuance. We are agreed, as I understand it, that the extension, generally, of roving wiretap authority to intelligence investigations is wise, and our disagreements, such as they are, exist at the margins.

It is useful, before addressing those disagreements, to recall the genesis of the roving wiretap rules and the fundamental reasons why they were extended to intelligence investigations. The Fourth Amendment requires that search warrants specify with particularity the place to be searched. This is intended to prevent the accidental or abusive search of an innocent person with, for example, a warrant obtained to search the home of another. As originally applied to electronic surveillance, the particularity requirement meant that law enforcement officers had to specify the particular phone they were intercepting.

Roving wiretap authority is a response to changing technology. Our original electronic surveillance laws stem from a time when phones were fixed in one place and linked to a network by a hard copper wire. Today, when phones can cross state and international boundaries at the speed of flight and where they are disconnected from any physical network, that model is antiquated.

In response to these changes in technology, in 1986 Congress authorized a relaxation of the particularity requirement for the investigation of drug offenses. Under the modified law, the authority to intercept an individual's electronic communication was tied only to the individual who was the suspect of criminal activity (and who was attempting to "thwart' surveillance by, for example, changing phones or locations frequently) rather than to a particular communications device. In 1998, Congress altered the standards somewhat to permit use of a roving wiretap when the target's conduct in changing telephones or facilities had the effect of thwarting the surveillance. Prior to September 11, it was not clear that theses authorities could be used to track terrorists, independent of predication to believe a crime was being planned. To clarify the law, and to close the potential gap, section 206 authorized similar techniques for foreign intelligence investigations—an extension that, on its face, seems reasonable.

My colleague agrees generally, but proposes two modifications of existing law. He begins, however, from a premise with which I disagree -- that the relaxation of the particularity requirement is constitutionally suspect. That premise influences, I suspect, much of the analysis: but it is, I think, not well founded. To be sure, no court has addressed the constitutionality of section 206, and the Supreme Court has yet to pass on the constitutionality of roving wiretaps generally. But it isn't as if the courts have never addressed roving wiretaps at all—indeed, every appellate court to consider the constitutionality of roving wiretaps has held that they do not violate the particularity requirements of the Constitution, reflecting a contextual understanding that these investigative practices are reasonable.

Reflecting this constitutional hesitancy, the opening essay suggests two modifications to existing requirements—an ascertainment requirement and a heightened identification requirement. Both seem unnecessary and unwise.

The imposition of an ascertainment requirement on law enforcement and intelligence agents would burden their ability to monitor the activities of suspected terrorists and, at the margin, decrease the utility of section 206. Where opponents see the prospect of the interception of innocent conversations, proponents are concerned that during the delay while an ascertainment is being made, or in circumstances where an ascertainment is uncertain, vital terrorism intelligence will be lost. The question then is the balance of risks and benefits. While the balance struck by the ascertainment requirement may make sense in the traditional criminal context, it makes less sense in the context of terrorism investigations.

To begin with, the statute has substantial safeguards against misconduct already—no interception may be authorized unless there is probable cause to believe that the target of the surveillance is a foreign power or an agent of a foreign power. There must also be probable cause to believe that "each of the facilities or places at which the surveillance is directed is being used, or about to be used,' by the foreign agent who is the target of the surveillance.

These requirements are subject to both administrative and judicial scrutiny prior to authorization, limiting substantially the chances that a "fishing expedition' of innocents will be initiated. In addition, a minimization requirement mandates the termination of surveillance upon the determination that an intercepted conversation is innocent. Given these protections there is little that the ascertainment requirement adds to the analysis and much harm it can do in creating hesitancy in our terrorism investigations.

But the ascertainment requirement is wrong for a more fundamental reason: It imposes a narrow law enforcement paradigm on the efforts to combat terrorism. The traditional law enforcement model is highly protective of civil liberty in preference to physical security. The post-September 11 world changes this calculus, principally by changing the costs from a mistake. Whatever the costs of failing to collect information regarding organized crime boss John Gotti might be, they are considerably less than the potentially horrific costs of failing to stop the next al-Qaeda assault. Thus, the theoretical rights-protective construct under which our law enforcement system operates must, of necessity, be modified to meet the new reality. We simply cannot afford a rule that "better 10 terrorists go unscrutinized than that one innocent be mistakenly subject to surveillance.' Asserting a direct equivalence between terrorist investigations and the traditional law enforcement construct—which is what the ascertainment requirement does -- misses this point altogether.

Consider next the identification requirement. Under the Patriot Act, agents may seek authority for an interception even when the identity of the suspect is not known (so long as probable cause exists to believe the person involved was an agent of a foreign power). Imposing a more substantial identification requirement would change that regime. If adopted, it would require agents seeking authority for a wiretap to specify the identity of the target and, if they were unable to do so, to describe with specificity the nature and location of the places where the interception would occur. In other words, in certain circumstances, intelligence agents might be unable to secure a warrant to conduct electronic surveillance because of the indefiniteness of their information.

The proposed modification of section 206, I submit, misses the point. To the extent that it calls for specificity with respect to the precise location or facility where the communication is occurring, it is a non sequitur. Government agents use roving wiretaps only when the location or facility where the communication is occurring is not known with precision--for the simple reason that those under surveillance are attempting to thwart surveillance by constantly changing their location and means of communication. They may have probable cause about a type or class of communications, but lack a particular description of the next "throw away' cell phone that will be used.

The alternate proposal to require that the individual who is the subject of scrutiny be precisely identified is equally unwise. In a domestic investigation, the identity of the suspect under scrutiny may often be well known, though drug dealers do, of course, use aliases. The problem becomes substantially more acute in the shadowy world of espionage and terrorism, where the identity of the investigative subject is often obscured behind a gauze of deceit.

Terrorists change their identity with frequency and often pose as other, real-world individuals. Often, the only description that the intelligence agency will be able to provide to identify the suspect is an alias (or several aliases). Sometimes the description of the terrorism suspect may be nothing more than a physical description. And, on still other occasions, it may consist only of a pattern of behavior (i.e., the person who regularly uses this series of phones, in this order, every third day). To insist that intelligence and law enforcement agents precisely identify the individual under scrutiny or the facility he will be using is, in effect, to limit the use of roving wiretaps in terrorism investigations.

Finally, it bears noting there is no practical necessity for the proposed revisions. Though section 206 has been the law of the land for more than three years, there have be no reported instances of abuse of this authority—despite scrutiny by the courts, Congress, and the Department of Justice's Inspector General. Apparently, the Department and the courts are applying the laws in ways that constrain discretion adequately. Whatever else may be said about the Patriot Act, even its most ardent critics must admit that they are basing their legislative proposals on the fear of potential abuse rather than reality of actual abuse.

James X. Dempsey Responds

Courts have upheld the constitutionality of roving taps in the criminal context precisely because those roving taps have the guiding standards that are missing from Section 206.

The Constitution requires all search warrants (a wiretap order is basically a search warrant) to "particularly describe the place to be searched.' Recognizing the crafty mobility of terrorists and other criminals, courts have held that this requirement can be satisfied by describing with particularity either the place (that is, the phone number or Internet account) to be monitored or the person to be monitored. The problem with Section 206 of the PATRIOT Act, as modified by the 2002 intelligence authorization act, is that it allows the issuing of roving tap orders without describing either the place or the person. While this leaves roving tap authority for national security cases on shaky ground, it can be clarified without tying the hands of intelligence agents.

The prior essay, by arguing that criminal standards are inapplicable to terrorism investigations, confuses the principles applicable at the trial stage of a criminal case with the much looser rules applicable at the investigative stage. In criminal and intelligence investigations alike, the law allows the government to cast a very broad net. Proponents of fixing the PATRIOT Act's surveillance provisions are not proposing standards suited to criminal trials. We are saying that bedrock constitutional principles must be followed in both criminal and intelligence investigations, and the particularity requirement is one such bedrock.

Another important protection is the so-called ascertainment requirement. The purpose of roving taps is to follow the bad guy, so unless the bad guy is being followed, the roving tap cannot and should not be activated.

The prior essay presents a false choice between fighting terrorism, on the one hand, and preserving the civil liberties of innocent people on the other. True, we have checks and balances in part to protect the innocent. But we impose limits on government agents also in order to focus their activities and make them more effective. The ascertainment requirement serves to focus government surveillance resources—it ensures that the FBI is not going up on a surveillance without reason to believe that the targeted person is using a particular phone or computer. In this way, the ascertainment requirement aids the fight against terrorism.

The prior essay does not really explain why ascertainment would be burdensome. If an officer is required to end surveillance after determining that the wiretap is intercepting the communication of innocent persons, it hardly seems onerous to require the officer to determine that the target is using the communication device before activating the wiretap in the first instance.

Without additional safeguards, Section 206 roving tap orders are little different from the "general warrants' that the Fourth Amendment prohibits.

Paul Rosenzweig The Last Word?

There is no false choice between preserving civil liberty and fighting terrorism. To the contrary, the right answer is to seek to maximize both values to the extent it is possible. In that regard, the ascertainment provision championed by the competing essay is one that misses the mark.

First, the ascertainment requirement in criminal law (18 USC § 2518(12)) applies only to "oral communications' (i.e. those revealed by a hidden microphone) and does not apply to the interception of "wire' or "electronic' communications, like the telephonic conversations subject to interception under section 206. Thus, the ascertainment requirement advanced as a response to the "problem' of roving wiretaps takes a rule used in criminal cases only for oral conversations (like those of Mafia members in an eating club) and applies it to telephonic and electronic communications. But this means that if the ascertainment requirement were adopted it would be something totally new to electronic interceptions and actually make it more difficult to intercept the telephonic communications of terrorists than of drug dealers—surely not the ideal answer.

Of equal importance, the companion essay does little to demonstrate the necessity of the ascertainment requirement in practical terms. Relying on the robust minimization requirements that continue to exist, the essay posits that ascertainment prior to use is just as easy as minimization once use begins.

But the reality is far different—minimization is based upon knowledge of ongoing conversations, while ascertainment is based upon suppositions regarding future events. It is absolutely predictable that the uncertainty of ascertainment will cause hesitancy in the initiation of an interception. And through the gap created by that hesitancy will flow terrorist communi-cations. Thus, given the probable cause requirements for a section 206 warrant and the minimization protections, the benefits of the ascertainment rule are modest at best, while the costs are likely to be quite real.