The ODNI’s New and Disappointing Prepublication Review Process

Editor's Note: This piece is cross-posted with Just Security.

On July 12, Director of National Intelligence Avril Haines signed Intelligence Community Directive 711, which establishes new requirements for intelligence community (IC) administration of prepublication reviews. It is the first IC-wide guidance on prepublication review, and it will lead to changes in the much-criticized (including by us) system of prior restraint on speech that applies to millions of current and former U.S. government employees. The directive includes a few of the reforms we have proposed and in some procedural respects represents small steps in the direction of fixing a badly broken system. But it also contains very disappointing speech-harming provisions, and even the potential improvements will depend on how individual intelligence elements implement the directive. On the whole, the directive is very disappointing.

Background on Prepublication Review

The U.S. government requires prepublication review for written works by certain current and former U.S. government officials. Those subject to prepublication review, including most former employees who have or had top secret/sensitive compartmented information (TS/SCI) clearance (us included), are required to submit prior to publication a vaguely defined and often sweeping range of material to the agency that issued the security clearance, which is typically the agency for which one worked. What precisely must be submitted differs from agency to agency (at present, each agency has its own policy). Many agencies, in their policies or as a matter of practice, require prepublication review of any material, whether fiction or nonfiction, that might be related to an employee’s government service. And the obligation applies for the rest of one’s life.

We have previously written about the problems with this system, which we have argued constitutes the largest system of prior restraint on speech in U.S. history. In particular, we have criticized the preclearance system’s lack of clear and reasonable standards or adequate process, and the chilling effect the attendant prior restraint has on the speech of millions of former government employees. And we have offered specific proposals for how to improve it.

We made many of these arguments in an amicus brief we submitted to the U.S. Supreme Court in support of the unsuccessful petition for certiorari in Edgar v. Haines. The government’s reply brief in opposition, filed with the Court in April 2022, noted that the House and Senate intelligence committees had called in 2017 for the director of national intelligence to issue “‘an IC-wide policy regarding pre-publication review’ with ‘improve[ments] to better incentivize compliance and to ensure that personnel fulfill their commitments.’” The brief stated that the Office of the Director of National Intelligence (ODNI) “has informed this Office that its review of IC prepublication-review policies is ongoing and that it is in the process of responding to that congressional request.” More than two years after the Court denied certiorari, this intelligence community directive is the IC’s response.

Limited Uniformity

Existing prepublication review policies are highly decentralized. Each agency has established its own system with its own procedures and its own rules about what must be precleared. That has meant that someone who had TS/SCI clearance to work at the Department of Defense may, for example, have to submit materials for review that someone with similar clearance at the State Department would not. (Some officials have clearances through multiple agencies, which further complicates the process.) Review of similar documents may take two weeks at one agency and two months at another. A piece of writing may be easily cleared by one agency and held up for review or cleared only with specific revisions by another.

The new directive “applies to the IC, as defined by the National Security Act of 1947” and to other elements that may be designated by the president as an element of the IC. The stated aim of the directive is to establish “uniform prepublication review rules among IC elements” (D.2). To do this, the directive requires IC elements to “establish a prepublication review process or modify their current prepublication process in a manner that is consistent with [the directive], including by creating or maintaining a public-facing website” (D.4). This should provide greater uniformity to a process that has been riddled with inconsistencies.

The directive only applies going forward, however. It “does not alter the responsibilities or obligations of individuals under law, federal regulation, or contract, including existing U.S. Government nondisclosure agreements (NDAs)” (D.16). That said, the directive provides that, “[u]pon request of covered personnel, IC elements shall afford covered personnel who have signed NDAs that predate this ICD an opportunity to sign new NDAs consistent with this ICD and cancel the prior NDAs.”

The uniformity imposed by the directive is limited. The directive by its terms “applies to the IC, as defined by the National Security Act of 1947” and to other entities so designated by the president. It defines “covered personnel” to include those “whose access to covered intelligence was or is sponsored by an element of the IC,” including those located outside the IC (C.2). Another part of Directive 711 states:

Current or former personnel who were not employed by an IC element (e.g., congressional staff or industry personnel) shall contact the employing entity’s security office, or the Defense Counterintelligence and Security Agency for industry personnel, to determine which IC element sponsored their clearance, and shall submit the prepublication review requests to the IC element or department that sponsored their clearance. [D.3.c]

These provisions seem to mean that anyone whose access to covered intelligence was sponsored by an element of the IC is subject to the directive. But the directive does not appear to apply to those with access to classified information if their clearance was sponsored by an element outside the IC.

It thus seems doubtful that the directive applies to executive branch officials whose security clearances were sponsored by a non-IC element of the Defense Department, the State Department, or likely many other departments. And the extent to which the directive applies to White House staff (such as the members of the National Security Council) is unclear. It likely applies to the many White House officials (especially on the National Security Council) detailed from IC agencies who received their clearances from those IC agencies. But it seems not to apply to the many members of the White House staff who are not detailees from IC agencies and who did not receive clearances from an IC element. (A prominent example of this gap would be former National Security Adviser John Bolton, who got caught up in a highly dysfunctional preclearance review process entirely within the White House.)

This means the government will continue to apply inconsistent prepublication review standards even after this directive, though there will be greater uniformity within the IC. This non-uniformity is likely not fixable by the director of national intelligence (whose jurisdiction is basically limited to the IC). The president could close the loophole with an executive order akin to Directive 711 that applies to all persons whose clearances are not sponsored by an IC element. Unfortunately, we cannot recommend that course of action because the directive has too many shortcomings as currently written to justify its extension.

A Disappointingly Broad Duty to Preclear

Directive 711 provides that it applies to:

non-official material that a fully-informed person granted access to covered intelligence might reasonably deem to contain or be derived from covered intelligence and that is intended for disclosure to, or discussion with, one or more individuals not authorized to access covered intelligence, or which is intended for public dissemination. [C.3]

We fear that this strangely worded passage imposes a novel and super-broad speech-harming duty to preclear.

The passage recognizes an important fact about the preclearance review process: A piece of information in a publication may appear to not contain or be derived from classified information based on the information set of the current or former employee and yet concern classified information when viewed in the context of the wide range of other materials unknown to the employee. There is no way that the vast majority of employees and former employees can know that broader context when deciding whether to submit material for prepublication review. If former employees are confident based on their information set that the material they wish to publish does not contain and was not derived from classified information, how could they know whether someone with “full” information might reasonably reach a different conclusion? They cannot.

And yet that is what the directive appears to require those covered by the directive to do. The duty to preclear is tied to what a “fully-informed person granted access to covered intelligence” knows. Past or present employees who are confident based on their government experience and general knowledge that a publication contains no classified information have, it seems, a duty to preclear unless they are also confident that a “fully-informed person granted access to covered intelligence” would also conclude that the writing contains no classified information. This is literally an impossible standard to satisfy—no partially informed person can know what the fully informed person might reasonably conclude. The directive therefore seems to require those covered by it to preclear practically every publication with any conceivable connection to government work or even a government program on which the person did not work, even if there is no reason for the person to think that the publication contains or derives from classified information. And, relatedly, it appears to give the government a basis to require preclearance of practically any writing it likes.

In short, ODNI has opted for a regime that gives absolute priority to security over speech, and that maximizes the IC’s opportunity to see in advance every publication by every person who has ever held a TS/SCI clearance to ensure that the publication does not contain classified information. The security maximalism of the “fully-informed person” standard denies any role for First Amendment values. Moreover, it ignores that excessive restrictions might, in fact, jeopardize security. Indeed, it is far from clear that the directive will achieve its aims. If the massively overbroad new standard is taken literally, the IC will be swamped by a resource-stretching avalanche of preclearance requests for publications that raise no conceivable security concerns. But since the standard is impossible to satisfy if taken literally, many employees will surely put their own glosses on it or decline to comply, possibly to the detriment of national security.

It is conceivable that the ultimate aim of Directive 711 is to encourage employees to preclear if they are not sure if a publication contains classified information. That would be a more reasonable standard. But that is not what the language says. If that is what the directive nonetheless aimed to achieve, ODNI should clarify the relevant preclearance standard. As it is, the directive embraces an extreme duty to preclear for the entire IC that will chill likely speech to an unprecedented degree.

Other Elements of Breadth in the Prepublication Review Obligation

Directive 711 imposes a broad prepublication duty in other respects as well.

The directive’s demanding objective standard appears to require preclearance for a current or former employee who wishes to “publish” information that appears in classified documents but is also widely publicly known, even if not officially acknowledged. In practice, this means that those who worked on matters that have been widely reported cannot speak about matters that everyone already knows, disabling their participation in public discourse. (The problem is exacerbated by the fact that, as the litigation over the targeted killing policy in the Obama administration shows, what constitutes “official acknowledgement” is not always clear.) If one accepts that a former employee’s implicit confirmation of widely known but not officially acknowledged classified information could harm national security, one may think this is a reasonable restriction. The two of us have differing views on that question. But whatever one’s views, there is no doubt that this is a significant implication of the new directive.

The directive also applies to a vast array of materials that includes but is not limited to “works of fiction or non-fiction, such as: prepared oral statements, resumes, editorials, videos, books, scholarly papers, scripts, screenplays, interviews, speeches, newsletters, websites, social media postings, brochures, graphics, briefings, articles, presentations, book reviews, or remarks for panel discussions” (C.3.b). In short, it is not limited to nonfiction materials. Nor is it limited to works intended for “publication” in the traditional sense, as it applies to, among other things, “remarks for … discussions.” This broad range of covered materials is not new, but it is noteworthy.

Prior Restraint of Members of Congress

As noted above, Directive 711 provides that:

[c]urrent or former personnel who were not employed by an IC element (e.g., congressional staff or industry personnel) shall contact the employing entity's security office, or the Defense Counterintelligence and Security Agency for industry personnel, to determine which IC element sponsored their clearance, and shall submit the prepublication review requests to the IC element or department that sponsored their clearance. [D.3.c]

Some commentators maintain that members of Congress are granted access to classified material “by virtue of the constitutional offices they hold.” But even those who advocate this view acknowledge that the access does not entitle members to sensitive compartmented information. Members are granted access to specific compartments by the executive branch based on a need to know a particular category of information. Congressional staff must apply for both TS/SCI security clearance and access to specific compartments. A condition of access to the information contained in those compartments, whether by members or staffers, is the signing of a non-disclosure agreement (NDA) that can, the directive makes clear, subject them to lifelong prepublication review obligations.

“[P]ersonnel … not employed by an IC element,” the directive explains, are subject to prepublication review through a process established by the element that provided the security clearance. Members and staff have long been signing NDAs as conditions of access to TS/SCI information that likely have contained similar obligations. Nonetheless, the implications of imposing on members and staff an expansive obligation to preclear “publications” through the IC for a lifetime are significant. Recall that the directive applies to “non-official material” that is the basis for mere “discussion” of the material. So if a member of Congress who had access to TS/SCI materials wants to discuss related matters, that discussion is now clearly subject to prepublication review by the agency that granted classified access to the information. If, for example, members wanted to criticize the U.S. program of torture at CIA black sites—a subject of significant disagreement between the CIA and Senate—they must submit their remarks to the CIA for clearance in advance. (If the remarks are made on the floor of the House or Senate, they would be protected by the Speech or Debate Clause, but that protection does not apply more broadly.)

On the Slightly Brighter Side

The new directive does contain improvements.

It states that the “purpose of prepublication review is to ensure the protection of classified government information, including intelligence and intelligence-related information, that upon disclosure reasonably could be expected to cause damage to the national security” (D.1). It also provides that the review should be “impartial, and free from political or policy bias” (D.5). These commitments should go without saying, but there is plenty of anecdotal evidence that the prepublication review process has been used for suppression of information that is not classified but that the government simply does not want made public. The tussle over the clearance of John Bolton’s book, for example, seemed to show that the Trump administration was holding up clearance at least in part because the material in the book cast Trump and his administration in an unfavorable light. Whether the prepublication review process will in fact be less susceptible to such viewpoint discrimination is, however, unclear, given that agencies remain in full control of their own review processes.

The directive also contains procedural improvements. It requires all IC elements to create a public web page for their prepublication reviews, give prompt acknowledgement of the receipt of a prepublication review request, and provide notice if the material has to be referred to another element. The directive also prohibits showing submitted material to those not involved in the review. One of us experienced this problem firsthand, discovering that a submitted book manuscript had been circulated far beyond those involved in the review of the manuscript for possible disclosure of classified information.

The directive requires each element to establish an appeals process. Up to this point, there has been no such process. The only option available to someone who was improperly refused permission to publish was to sue the government or proceed with publication and accept the consequences that might follow. The directive says that the submitter has 60 business days to appeal, that the IC element “shall strive to provide a final determination within 30 business days from receipt of the appeal” (D.10), and that the IC element must establish an “Appellate Authority” to handle such appeals (D.12). Beyond that, the directive does not specify how the appeals process should work or the standard of review that should be applied, and it does not bind the IC to any time limitation to resolve the appeal. Like much else in the directive, therefore, the effectiveness of the appeals process will depend entirely on how it is implemented in the agencies—which means that the end result is likely to be uneven at best.

Further Reform

There are several opportunities for further reform beyond those already mentioned.

First, the government should consider a temporal limit on who is subject to prepublication review. The new directive still applies to millions of former U.S. government employees who have not served for a decade or more, and whose access to classified information is so dated as to be practically irrelevant. It is a huge waste of the government’s time to review most materials written by such people and a huge burden on their free expression to mandate preclearance in this context.

We favor limiting the scope of those subject to prepublication review to those whose TS/SCI clearance has been active within the last 10 years, perhaps subject to discrete and targeted exceptions for certain super-sensitive programs. This would allow the government to focus on those who most likely possess information that could do harm to U.S. national security, justifying the prior restraint on their speech. It is important to understand that limiting the temporal scope of those obligated to submit to prepublication review would not mean that those without a prepublication duty could publish classified information. Any unauthorized disclosure of classified information is a crime, subject to prosecution under a number of federal statutes, including the Espionage Act. Prepublication review is an exceptional ex ante prophylactic measure whose justification weakens over time.

An alternative would be to make prepublication review optional for those who have not had access to classified information for a decade or more but offer them a safe harbor if they choose to submit material for review. Under this approach, they would no longer be obligated to submit materials for prepublication review, but they would be permitted to do so. If they voluntarily submitted material, that material, once cleared, could not be the basis for any future criminal, civil, or administrative punishment.

Second, the new directive is clearer that prepublication review should be carried out expeditiously, but it does not provide any binding limits. It provides that short, time-sensitive requests “shall be handled as expeditiously as practicable” (D.4.a) but does not specify what that means. It also states that if any review exceeds 30 business days, “the receiving IC element shall immediately notify the submitter with a status update and reason for the delay and then provide an update every four weeks thereafter” (D.4.b). This is an improvement over the status quo, but it is unlikely to prevent endless delays—a problem that has plagued the current system. In Freedman v. Maryland, the Supreme Court held that prior restraints must yield a determination “within a specified brief period.” The directive should be updated to include binding outer time limits, so that former employees are not left at the mercy of a process that can drag on indefinitely. Creating certainty would also incentivize better participation. As it is, many of those subject to prepublication review simply ignore the requirement because it is so cumbersome and unpredictable.

Third, the creation of an appeals process is a step forward, but leaving it up to the agencies to design and manage the appeals process makes the step a very small one. One stated goal of the directive is to create greater consistency across the many IC units. At a minimum, the directive should have provided more uniform standards—including the standard of proof and firm time limits for resolution of the appeal. For some matters, it may have been possible to create a centralized appeals process, though a central appeals body might have to rely on specialized expertise located within the agencies. As written, the directive will be interpreted differently by the many different IC elements, without any mechanism to identify or resolve those inconsistencies.

Conclusion

Directive 711 contains modest procedural improvements. But the core preclearance standard is exceptionally broad, and the efficacy of the core procedural improvement—an appeals process—will turn on the almost-certainly uneven implementation of the directive by the agencies. Considered in the round, the new directive is a missed opportunity.