Intelligence Surveillance & Privacy

A Brief History of Programmatic Collection Pre-Section 702

James Petrila
Wednesday, April 12, 2023, 8:16 AM

Given the importance of Section 702’s collection authorities for national security, it is worth reviewing how it compares to the original FISA.

National Security Agency Headquarters in Fort Meade, Maryland, 2013. (Trevor Paglen,; CC0 1.0,

Published by The Lawfare Institute
in Cooperation With

Section 702 of the Foreign Intelligence Surveillance Act (FISA) will expire at the end of this year unless renewed by Congress. This section of FISA is part of the critical foundation for the intelligence community’s ability to collect foreign intelligence on a range of significant topics important to national security on a programmatic basis. It has been controversial since its predecessor, the Protect America Act, was first passed in 2007 in response to public disclosures of President George W. Bush’s Terrorist Surveillance Program, which is perhaps better known by its code name Stellar Wind. Section 702 in its current form was enacted as part of the FISA Amendments Act in 2008. Unlike most sections of FISA, Section 702 is subject to periodic renewal. It was renewed in 2012 and then again in 2017, with broad bipartisan support. While Section 702 has always drawn opposition from civil libertarians and privacy advocates in both the United States and the European Union, the significant danger this time around is the opposition among certain members of the narrow House Republican majority. 

In a nutshell, Section 702 authorizes the intelligence community to target non-U.S. persons located overseas for the purpose of collecting foreign intelligence information within certain defined parameters: the collection occurs in the U.S.; the purpose of the collection, but not the individual target, is within an area authorized by the Foreign Intelligence Surveillance Court (FISC); and the retention and dissemination of the information is consistent with FISC-approved minimization procedures as well as FISC-approved procedures and guidelines for targeting and querying. Given the importance of Section 702’s collection authorities for national security, a review of the history of the original FISA as passed in 1978 demonstrates that Section 702’s authorities are consistent with the act’s original intent and are the result of belated efforts by Congress (and even more belated efforts by the executive branch) to ensure that FISA’s collection authorities reflect the revolution in communications technology that has occurred since FISA was first passed in 1978.

The need for something like FISA was made clear as early as 1972, when the Supreme Court ruled in the famous Keith Case that electronic collection in domestic security cases fell within the Fourth Amendment’s warrant requirement as set forth in Katz v. United States. The Court made clear that its decision did not address the scope of the president’s surveillance power regarding the activities of foreign powers either inside or outside the U.S. Subsequent congressional investigations by the Church and Pike committees, however, revealed that the National Security Agency (NSA) had engaged in substantial unauthorized warrantless collection directed at U.S. persons in the U.S. It was not until 1978 that Congress was able to pass a statutory fix to address these reported abuses. 

To understand why Section 702 is consistent with FISA’s original intent, it is helpful to review what communications Congress intended for FISA to cover and what communications were deliberately excluded from FISA’s coverage. Congress carefully crafted language in the definitions section of FISA to ensure that “electronic surveillance” specifically excluded the NSA’s satellite collection program. The statutory language was intended to permit the NSA to continue with its robust satellite collection program directed against non-U.S. persons located overseas, even if the collection occurred in the U.S. Why exclude satellite collection? Because in 1978, international communications of foreign intelligence interest were carried almost exclusively by the INTELSAT constellation of commercial satellites. The original FISA also recognized that this type of programmatic intelligence collection, which the NSA had engaged in since the agency’s inception, was not susceptible to a warrant requirement because of the nature of the target set. In other words, the drafters of the original FISA wanted to ensure that the intelligence community continued to have access to a vast array of communications carried by commercial satellites where the target was a non-U.S. person located overseas even if that meant that a considerable amount of U.S.-person information would be incidentally collected in the process. 

Furthermore, that information—to include incidentally collected U.S.-person information—would be collected under the president’s Article II authority without any judicial review, to include by the FISC, so that the collectors at the NSA would have the necessary flexibility and agility in their collection activities, subject to minimization procedures (to include limitations on how the databases could be queried) that would protect the privacy interests of U.S. persons.

The method by which the original FISA achieved this goal is apparent in the definitions section of FISA, found at 50 U.S.C. § 1801(f)(1)-(4). FISA had four definitions of “electronic surveillance[.]” If a collection method met one of those four definitions, then the collection was covered by FISA and the collection would require an order issued by the FISC before collection could occur. If the collection method did not meet one of those four definitions, then the collection was not covered by FISA, which meant that it could occur exclusively under the president’s Article II authority subject to the “reasonableness” requirement of the Fourth Amendment.

These definitions of “electronic surveillance” are referred to in this article as (f)(1) through (f)(4). The (f)(4) definition, which covers “information, other than from a wire or radio communication,” in the original FISA was intended to encompass close access programs. The (f)(4) definition is relevant to Section 702 for those cases where collection is directed at non-U.S. persons located abroad whose only ties to the U.S. are emails of foreign intelligence interest stored in the U.S. The first three definitions are the most important for understanding continuity between the original FISA and Section 702, in part because Section 702 relies on the definitions of the original FISA. These definitions of “electronic surveillance” divide the world of communications into two parts: wire communications and radio communications. These definitions further divide the world into two parts: the U.S. and the rest of the world. “Radio communication” in 1978 generally meant communications carried via satellite. “Wire communication” meant “any communication while it is being carried by a wire, cable, or other like connection” operated by a common carrier (emphasis added). In general, radio communications could be collected by placing a satellite dish in the footprint of the satellite carrying the targeted communications. Wire communications required a traditional wiretap, generally implemented at a central office switch controlled by the local common carrier. FISA provided the means to compel cooperation of the common carrier by means of an order issued by the FISC. 

We’ve already looked at the definition in 50 U.S.C. § 1801 (f)(4), but in order to understand the original FISA more fully, we should first delve into the definitions of electronic surveillance, as laid out in 50 U.S.C. § 1801 (f)(1)-(3).

The (f)(1) definition requires the government to obtain a FISA warrant to collect any communication, either wire or radio, to or from a known U.S. person in the U.S. where that person is the target of the collection and where the collection occurs under circumstances in which a person would have “a reasonable expectation of privacy and a warrant would be required for law enforcement purposes.” The importance of the (f)(1) definition is that it makes clear that any electronic surveillance directed at a known U.S. person in the U.S. would be covered by FISA’s warrant requirement.

The (f)(2) definition covers the nonconsensual acquisition of “any wire communication to or from a person in the United States” where the acquisition occurs in the U.S. (emphasis added). Note that (f)(2) covers more than U.S. persons in the U.S. and that the “reasonable expectation of privacy” language is not included in (f)(2)’s definition. This definition made it clear that any collection against foreign embassies located in the U.S. would be covered by FISA and that a FISA order would be required to collect against embassies in the U.S.

Finally, the (f)(3) definition covers the acquisition of radio communications where both the sender and all intended recipients are located within the U.S. if the collection occurs under circumstances in which a person would have a reasonable expectation of privacy and a warrant would be required for law enforcement purposes.

While collection of satellite communications fell outside of FISA’s definition of “electronic surveillance” so long as the target of the collection was outside the U.S., internal Department of Defense regulations governed the way in which the NSA could retain and disseminate U.S.-person information that was incidentally collected outside of FISA’s coverage. The artful drafting of the (f)(3) definition allowed the NSA to continue its warrantless collection of international communications on a programmatic basis while the (f)(1) definition prohibited the NSA from targeting known persons in the U.S.

This system met the intelligence community’s needs for the first decade of FISA’s existence. So long as the NSA had satellite dishes that were within the footprint of the INTELSAT constellation that carried the bulk of the world’s international communications, the technology to sort through the mass of collected communications, and the Department of Justice approved minimization procedures, life at Fort Meade must have been pretty good. The NSA’s ability to collect foreign signals intelligence (SIGINT) on topics of critical national security and foreign policy interest on a programmatic basis enabled government efforts in the areas of counterproliferation, counternarcotics, counterterrorism, counterintelligence, and foreign relations, to name several likely foreign intelligence collection topics. The problem, though, is that technology is not static. The development of fiber optic cables, and the successful deployment in 1988 by a consortium of AT&T, France Telecom, and British Telecom of Transatlantic Cable Number 8, or TAT-8, the first undersea fiber optic cable, caused a revolution in the way communications could be transmitted. 

The tandem development of the world wide web meant that the world was about to witness an explosion in sheer volume of communications. No longer would communicants need to rely on Western Union or AT&T to send their telexes via satellite. The world wide web made email available to anyone with a computer and an internet link. But these developments also meant that the definitional structure of FISA, with its critical distinction between satellite and wire communications, was about to become obsolete. If the purpose behind FISA’s disparate legal treatment of satellite and wire communications was to enable the NSA’s robust international communications collection program, then what would happen when, courtesy of fiber optics, the bird (satellite) suddenly became a duck (an undersea fiber optic cable)? 

The NSA’s legal authority to collect international communications in the U.S. suddenly was dependent on whether a smart switch at an international gateway facility (IGF) decided to send a communication via satellite or, more likely, via any one of an increasing number of undersea fiber optic cables. Within a few short years, international communications that touched the U.S. were more likely to be carried by wire than by satellite. By definition, any international wire communication, such as an email, that had one recipient in the U.S. suddenly became an (f)(2) wire communication rather than a “radio” communication that, courtesy of (f)(3), was outside of FISA’s coverage.

The NSA might have seen this problem developing in the early 1990s. And there was a clear statutory fix available—one that would have eliminated the distinction between wire and satellite communications in those situations where the target was a non-U.S. person located overseas (subject, of course, to minimization and other procedures). It does not appear, however, that there was ever a concerted effort to amend FISA to account for the revolutionary changes in technology that had upended FISA’s carefully crafted language. We can speculate as to why: The Soviet Union had collapsed, Congress was looking to slash spending as part of the post-Cold War “peace dividend,” and Sen. Daniel Patrick Moynihan (D-N.Y.) went on the Senate floor and speculated that the CIA was no longer needed and should be absorbed by the Department of State. In our pre-9/11 world, international terrorism was still viewed primarily as a law enforcement problem. President Clinton’s first campaign was characterized by the mantra “it’s the economy, stupid.” During this time, it was uncertain whether the NSA’s broaching of the topic of FISA amendments would result in a favorable outcome. The legal structure of FISA meant that the underlying question of collection was no longer as simple as identifying real estate in which to put a satellite dish to enable collection. Any discussion of enabling collection at the U.S. end of those high-value communications now migrating from satellite to fiber optic cable inevitably would require a discussion of the need to compel U.S. common carriers to cooperate with the NSA in its collection program, and how that should be done. Because of the agile and flexible targeting methodology that enabled programmatic collection, there certainly could not be a traditional FISA order identifying the specific communications and specific overseas targets that could be collected at the IGF.

The 9/11 attacks should have made it clear that FISA had a technology problem that required changes to the statute. Hypothetically, terrorist-related emails transmitted via fiber optic cable that had one addressee in the U.S. were covered by (f)(2) and could be collected only pursuant to a FISA order even when the target was an al-Qaeda operative plotting overseas. The same communication transmitted via satellite potentially could be collected. For example, in a communications stream between a European-based al-Qaeda cell that includes a person in the U.S. on the cc: line, the smart switch at the IGF could decide to send that communication via satellite. Assume that the NSA has the overseas cell on coverage and is able to collect the email in the U.S. (via the satellite downlink) because it is a satellite communication. Then, the NSA would share the incidentally collected information about the person in the U.S., consistent with minimization requirements, with the FBI. Now assume the same smart switch decides to send that same email via fiber optic cable. Prior to what is now Section 702, the NSA would have been unable to collect the very same communication, targeting the very same al-Qaeda cell, because of the mode of communication. Failure to renew Section 702 would re-create the same problem for the intelligence community, not only for counterterrorism, but for every other significant topic of foreign intelligence that currently is enabled by the programmatic collection authority of Section 702. 

The immediate aftermath of 9/11 would have been an excellent time for the Bush administration to seek an amendment to FISA, either as a stand-alone issue or together with the Authorization for the Use of Military Force (AUMF) that Congress passed shortly after the attacks. But the administration decided to take a different approach. Based on creative legal reasoning from the Department of Justice’s Office of Legal Counsel, the Bush administration instead launched Stellar Wind, which was a program in which the administration in essence created its own amendment to FISA that allowed it to ignore the specific language of FISA and treat satellite and wire communications collected in the U.S. as if they were indistinguishable. The administration also limited briefings on Stellar Wind to the so-called Gang of Eight. Published reports have described how the program was routinely renewed every 90 days by the president. In his book “A Higher Loyalty,” James Comey gives a dramatic account of how Stellar Wind became somewhat more constrained when he served as deputy attorney general in the George W. Bush administration. When the New York Times reported extensively on the program in a Dec. 16, 2005, article, congressional and public reaction was not positive. The Department of Justice attempted to explain itself in a Dec. 22, 2005, letter signed by the assistant attorney general for legislative affairs, but no amount of citing to the president’s Article II authority could overcome the fact that the Stellar Wind program appeared to conflict with the plain “electronic surveillance” language of FISA. The full Department of Justice inspector general investigation into Stellar Wind is available here

Once the Stellar Wind cat was out of the bag, the focus was on when and how Congress would act to amend FISA. The issue became even murkier when one FISC judge authorized the program to continue but a different FISC judge decided, shortly thereafter, that the program could not continue consistent with FISA as then drafted. After a temporary fix (the Protect America Act, passed in August 2007 as a short-term solution to allow the program to continue until a permanent solution could be found), Congress enacted FISA’s Section 702 as part of the FISA Amendment Act of 2008. Section 702 remains faithful to the original intent of FISA in that it allows targeting of non-U.S. persons located outside the U.S. to acquire foreign intelligence information on a programmatic basis. For purposes of Section 702, “foreign intelligence information” has the same definition as in the original FISA. Section 702 allows the attorney general and the director of national intelligence (DNI) to authorize jointly, for a period of up to a year, the acquisition of foreign intelligence based on the issuance of an order by the FISC. Collection under Section 702 (“acquisition” is the language used in the statute) may occur only in accordance with targeting and minimization procedures that are approved by the FISC, as well as with the submission of a joint certification by the DNI and the attorney general. Section 702 mandates that the FISC ensures that these certifications meet the statutory requirements set forth at 702(h), and that the targeting, minimization, and querying procedures and practices comply with the Fourth Amendment. 

The incorporation of FISC review and approval of attorney general and DNI certifications, as well as the inclusion of minimization and other procedures and guidelines, is an elegant solution for several reasons. First, Section 702 eliminates the distinction between modes of communication (which, in practice, had been largely irrelevant since the early 1990s) and focuses instead on the target of the collection. Given the issues associated with Stellar Wind, as well as the vast explosion of data made available by the internet, the drafters of Section 702 did not allow the executive branch to have exclusive control over programmatic collection. Congress provided that the FISC would play a significant role in reviewing not only the statutorily required procedures and guidelines but also the manner in which these procedures have been carried out. The DNI now releases redacted versions of the FISC’s Section 702 orders. The public release of these orders, even in redacted form, provides a level of transparency that did not exist for the NSA’s original programmatic collection. The public availability of these orders (for instance, the orders of Judge James Boasberg) provides an unprecedented level of transparency into the way in which Section 702 guidelines and procedures have been implemented. This publicly available information serves as a strong impetus for internal improvements in the use of Section 702 authorities; robust congressional oversight by the intelligence committees relies on the full classified report. In addition, Section 702 provides a statutory basis for congressional oversight, protections from liability for service providers, and a means of redress. Finally, while Executive Order 12333 sets forth the requirement to obtain attorney general authorization before the intelligence community may target a U.S. person located abroad, Section 702’s close relative, Section 704, gives the FISC jurisdiction to approve an application to target U.S. persons who are located abroad. These provisions of FISA taken as a whole constitute an elegant solution to the issues raised by programmatic collection in the age of the internet.

If Congress allows Section 702 to expire, it would mark a radical reduction in the ability of the intelligence community to collect critical foreign intelligence. It would also constitute a radical break from decades of authorized SIGINT activity that has allowed the intelligence community to provide policymakers with intelligence that matters on the broadest range of transnational issues (weapons proliferation, terrorism, international narcotics trafficking, and state and criminal cyberattacks) as well as significant intelligence directed against national security threats emanating from Russia, China, Iran, and North Korea. The opposition of civil libertarians is not surprising, as they generally have been skeptical of any broad authorities granted to the intelligence community, particularly in the area of technical collection. This skepticism can be positive when it focuses on ways to increase transparency and rigorous oversight. What is new is the overt hostility of a number of ultraconservative Republicans whose opposition to renewing Section 702 seems to be based largely on distrust of the FBI and the intelligence community’s reporting on Russia’s aggressive interference in the 2016 election on behalf of the Trump campaign.

If the Biden administration and bipartisan proponents of Section 702 want to address the concerns expressed by opponents of renewal, there are several steps they could take before Section 702 expires. A major expressed concern lies in the way in which the databases are queried once the information has been collected. These complaints are focused largely on the way in which the FBI seems regularly to have queried Section 702 databases for a variety of purposes not necessarily linked to foreign intelligence collection. It is important that these issues be dealt with separately. Failure to renew the programmatic collection authorities of Section 702 will result in a significant reduction in the collection of critical intelligence. Discussion of limitations on searches targeting incidentally collected U.S.-person information possibly contained in Section 702 databases is a separate issue that can be addressed apart from the underlying collection authorities. It is worth noting that recent reporting indicates that the FBI has significantly reduced its querying of databases that contain Section 702 data by taking the simple step of requiring that any general search of available databases by the FBI exclude Section 702 information. Congressional oversight of Section 702 is critical, and this is an area in which the oversight committees are actively engaged. Oversight of the ways in which the intelligence community uses the vast amounts of data collected under Section 702 should be aggressive, and changes to the current language of Section 702 may be necessary to ensure that databases are queried only for information related to foreign intelligence without a court order. Any decision to terminate this collection authority, however, would be a totally different matter. 

As has been argued recently in Lawfare, Section 702 is not without its problems in implementation. After all, any collection program that, by its nature, will result in the incidental collection of U.S.-person information will generate controversy and will require robust oversight. The failure to renew Section 702 collection authorities, however, would be a boon to the United States’s enemies around the world. In the absence of Section 702 authorities, the intelligence community’s collection authority would revert to the (f)(1) through (f)(3) authorities, which means that it would no longer be authorized to collect data in the U.S. that is transmitted on fiber optic cables. Programmatic collection, for the first time in well over half a century, would no longer exist in any meaningful sense. This would take the intelligence community back to the pre-9/11 era in terms of collection authority, leading to a significant loss of valuable foreign intelligence at a time of Russia’s war of aggression in Ukraine, a possible Chinese invasion of Taiwan, and Iran’s continuing effort to develop a nuclear weapon. Given today’s national security threats, it would constitute the intelligence equivalent of unilateral disarmament.

James Petrila had a thirty-year career as a lawyer at the National Security Agency and the Central Intelligence Agency. He also served as a Deputy Legal Advisor at the National Security Council from 2013-2015. He currently is an adjunct professor of law at George Washington University School of Law. He has a JD from University of Virginia School of Law, an MA in Russian History from Stanford University, and a BA in Russian Studies from Knox College.

Subscribe to Lawfare