Executive Order 12958, enacted in October, 1995, included several changes to the system by which government information is classified. In particular, the order mandates that many documents over 25 years old must be declassified by the year 2000. To the extent that this order accelerates the release of historically relevant information and discourages overclassification of new documents, it is a step in the right direction. However, further action should be taken both to expedite the process of declassification and to reduce the level of classification of some forms of information.

Progress towards automatic declassification threatens to be slowed because agencies expect to apply for the exemption of a large number of documents. The CIA, for instance, has declared that 64% of its information may be eligible for exemption. Also, the Information Security Oversight Office (ISOO), an office within the National Archives that has been given the responsibility to enforce some of the provisions within the executive order, has insufficient resources to carry out its duties.

This study recommends that the ISOO budget be significantly expanded, because a relatively small investment in the office would greatly improve ISOO's capability to oversee and assist the intelligence agencies in declassification efforts. Perhaps the additional funding for ISOO can be found by requiring intelligence agencies who have requested that a large portion of their documents over 25 years old be considered for exception to automatic declassification to lend financial support to the office. A threshold could be established, and agencies that apply for more exemptions than the threshold amount would have to contribute resources to ISOO. This would have the additional benefit of establishing an incentive to reduce the volume of requests.

The question of how the declassified information is to be distributed to the public is only addressed in general terms in executive order 12958, which mandates that a government-wide declassification database be established consisting of all declassified documents. How declassified information is distributed will affect the cost of the program and its effectiveness in the eyes of the public. However, distribution options that have been developed for the FOIA system provide hope for answering this question in a cost-effective way. The National Security Archives at George Washington University, for instance, is a privately funded non-profit organization that maintains a large collection of documents released through FOIA. The Archives maintains a public reading room and a database of FOIA requests, and it compiles collections of materials on popular topics on CD-ROM. The IC should seek advice from organizations such as these in order to develop an archive for declassified information that would be only partially funded by the government.

The justification for classifying new information must be reconsidered in light of recent changes in foreign policy and information technology. Classification of new information is costly for both industry and the government, which together spent $5.6 billion on the protection of classified information. Costs are incurred indirectly, as well, by making it more difficult to share information with organizations such as the United Nations during joint military operations. In the area of nuclear non-proliferation, the relaxation of security measures to allow other governments to inspect nuclear materials within the US has resulted in agreements to increase the security of nuclear weapons materials in the former Soviet Union. Finally, the increasing variety of national security threats has out paced the intelligence community's ability to analyze all of the information accumulated on those threats. For instance, in 1994, overhead imagery which was eventually released to the UN Security Council depicting mass graves in Bosnia was only discovered months after non-government organizations had investigated the incident. Had the images been available to a broader audience, they might have been discovered sooner and have had a more timely impact on US and world policy towards Bosnia. For these reasons, DCID 1/7, which focuses on producing intelligence in a format that is easily releasable to foreign governments and international organizations, should be extended to explicitly include intelligence for which further cooperation with organizations outside of the IC would be beneficial.

These areas include environmental intelligence, health intelligence, support for law enforcement, efforts to combat global organized crime. Although DCID 1/7 applies to all subjects of intelligence, the motivation conveyed in the memo accompanying it from DCI Deutch was support to military operations (SMO). The non-military subjects listed above should receive equal attention from the next DCI. In particular, the directive should include techniques analogous to tear lines for satellite imagery, such as reducing the resolution of the images, so that they may be easily distributed outside of the intelligence community if necessary. Also, the DCI should recommend that ISPAC, a council of individuals outside of the government who set priorities for declassification under E. O. 12958, include representatives from the fields listed above. In this way, the benefits expected for SMO under the directive can also be realized for non-traditional, but nevertheless important, intelligence objectives.

Making this statement a reality for the leaders of the nation has been the challenge of U.S. intelligence agencies for 50 years. There are many ways to describe the mission of the intelligence community (IC), but this one divides its mandate into two parts. First, intelligence officers must ensure, as best they can, that they "know the truth," a task which involves the collection of best possible information and its analysis in the best possible ways. Then, they must empower others to strengthen national security with the intelligence they produce, in effect making the nation free. Unfortunately, secrecy can be beneficial to the first task but detrimental to the second.

How can the need to protect the information within the IC be balanced with the need for other agencies and individuals to use it? Intelligence officers are continually faced with this question in a variety of specific forms. On one hand, the release of information about intelligence gathering capabilities increases the chance of those capabilities being undermined. On the other hand, excessive secrecy wastes resources and diminishes the value of the information by preventing its use by those who could benefit from it. Today, after the Cold War has ended and as the kinds of information available throughout the world are changing, an evaluation of the way that the IC answers that question is especially appropriate.

The most recent evaluation of the classification system culminated in 1995 with Executive Order 12958, which made several changes to the way information is classified and declassified within the Federal Government. One of the most controversial of these changes was the introduction of automatic declassification for documents over 25 years old. I argue below that this executive order was a step in the right direction, and, in particular, the automatic declassification program should be retained. In order to ensure its effective implementation, a branch of the National Archives known as the Information Security Oversight Office (ISOO) should be expanded with funds from those agencies that have applied for exemptions to automatic declassification for some of their archives, which ISOO must assist in processing.

More could be done to improve the classification of current documents, however, than what is found in the executive order. The IC should modify its classification procedures to provide for greater eventual coordination with other parts of the U.S. Government, foreign governments, and international organizations in areas such as environmental intelligence, health intelligence, and assistance to law enforcement. This could be accomplished by specifically mentioning these areas in the Director of Central Intelligence's directive to write at the lowest classification level possible, DCID 1/7. Also, specific techniques should be developed to facilitate the sharing of imagery intelligence that retain the value of the images to the consumer but also protect the secrecy of the capabilities of their sources. Changes such as these hold the promise of improving the effectiveness of coordination with outside organizations while maintaining secrecy at the minimum level necessary for national security.

CURRENT CLASSIFICATION SYSTEM

In April 1995, President Clinton issued Executive Order 12958, "Classified National Security Information," which established the current classification system used throughout the Executive Branch. The policy set out in the previous executive order, issued by President Reagan in 1982, had drawn criticism from non-governmental organizations such as the Society for Professional Journalists for being too secretive for the Post-Cold War era. Clinton's executive order was the result of an effort to reform the classification system that began with the formation of the Task Force on National Security Information in April, 1993. The final draft contains guidelines for which kinds of information may be classified, how it may be declassified, and which agencies have responsibility over the process.

The Classification of Information

Executive Order 12958 authorizes the classification of seven categories of information when its release poses an articulable threat to national security:

(a) military plans, weapons systems, or operations;
(b) foreign government information;
(c) intelligence activities (including special activities), intelligence sources or methods, or cryptology;
(d) foreign relations or foreign activities of the United States, including confidential sources;
(e) scientific, technological, or economic matters relating to the national security;
(f) United States Government programs for safeguarding nuclear materials or facilities; or
(g) vulnerabilities or capabilities of systems, installations, projects or plans relating to the national security.

Information that falls under one of these categories may be classified as "Top Secret", "Secret", or "Confidential", depending on the degree of damage that its release is expected to incur. The decision to classify information may only be made by "original classification authorities:" the President, agency officials designated by the President, and, except in the case of "Top Secret" information, designees of those who have been authorized by the President. Information is specifically prohibited from being classified in order to "(1) conceal violations of law, inefficiency, or administrative error; (2) prevent embarrassment to a person, organization, or agency; (3) restrain competition; or (4) prevent or delay the release of information that does not require protection in the interest of national security." It is also prohibited to classify scientific research that cannot be linked to national security and documents that have already been lawfully released to the public.

Executive Order 12958 introduced new classification procedures intended to simplify the eventual declassification of information. In order for a document to be classified, declassification instructions that specify a date or an event after which the document can be released must be specified. If a specific date or event cannot be determined, the default date for declassification is ten years after it was classified. Another new requirement for classifying a document is to include on the cover "a concise reason for classification which, at a minimum, cites the applicable classification catego ries in . . . this order." In contrast, the previous executive order permitted documents to be marked with the notation "Originating Agency's Determination Required" in lieu of more specific declassification instructions and did not require an explicit reason to be cited for the decision to classify. E.O. 12958 does, however, permit the reclassification of documents for successive periods up to ten years at a time, and exceptions can be made to the ten year classification limit for reasons similar to the exceptions to automatic declassification described below. The key difference between Clinton's policy and the previous one is that special attention must be paid to preserve a document's classified status after a period of time, as opposed to removing it.

Provisions for Declassification

Perhaps the most significant departure of President Clinton's order from the previous one is the mandate to automatically declassify all documents over 25 years old within five years of the effective date of the order. E. O. 12958 also continues the programs of systematic declassification and mandatory review upon public request that appeared in President Reagan's order.

Automatic declassification, as described within the executive order, is a misnomer in the sense that the name implies the release of information in bulk with little prior review. While this is the intent for much of the archives, exemptions can be made for information which could:

(1) reveal the identity of a confidential human source, or reveal informa tion about the application of an intelligence source or method, or reveal the identity of a human intelligence source when the unauthorized disclo sure of that source would clearly and demonstrably damage the national security interests of the United States;

(2) reveal information that would assist in the development or use of weapons of mass destruction;

(3) reveal information that would impair U.S. cryptologic systems or activities;

(4) reveal information that would impair the application of state of the art technology within a U.S. weapon system;

(5) reveal actual U.S. military war plans that remain in effect;

(6) reveal information that would seriously and demonstrably impair relations between the United States and a foreign government, or seriously and demonstrably undermine ongoing diplomatic activities of the United States;

(7) reveal information that would clearly and demonstrably impair the current ability of United States Government officials to protect the President, Vice President, and other officials for whom protection services, in the interest of national security, are authorized;

(8) reveal information that would seriously and demonstrably impair current national security emergency preparedness plans; or

(9) violate a statute, treaty, or international agreement.

The broad scope of these exemptions provides the opportunity to override automatic declassification often within the IC. In order for an exemption to be granted at the onset of the program, however, the agency must request it from the President. Requests that occur later in the process are handled by the Interagency Security Classification Appeals Panel (ISCAP). The President may deny the exemption or require the release of the information earlier than the requested date of declassification.

Under the order, Government agencies must also undertake systematic review of information with historical value but not eligible for automatic declassification. Material is chosen on the basis of its likelihood of being declassified upon review and recommen dations from the Information Security Policy Advisory Council (ISPAC), a group of seven individuals appointed by the President who are interested in declassification but are not employed by the Government. The agencies must also review information for declassification at the request of the public, a requirement referred to as "mandatory review," if the information has not been reviewed for two years and if the request is specific enough so that the information can easily be found.

Implementation of Executive Order 12958

There are tentative signs that E. O. 12958 has reduced the level of classified materials within the IC, but the prospects of meeting the goal set for automatic declassifi cation remain uncertain. According to Steven Garfinkel, Director of the Information Security Oversight Office, the rate of declassification has increased dramatically during the past couple of years:

This program [Executive Order 12958] has been in place now for a little over a year . . . , so we have a fair impression of what the last year has been. . . . One good thing has been that we have actually declassified a tremendous amount of material. I can't give you an exact number yet because we are a ways from reporting that to the President, but we are gathering data right now that indicates that -- and we're about to report that 1995 was an outstanding year -- 1996 looks like it is going to be even better in terms of the amount of material that was declassified.

ISOO's 1995 Report to the President was released after the interview; it states that the number of decisions to classify new documents decreased by over 20% in 1995 and that agencies declassified twice as many pages as they did in the year before through system atic review. As for progress in automatic declassification, John Carlin, Archivist of the United States, stated at a conference on declassification in December that:

Actually, for the most part, the results of the first year have been good. Most agencies have put declassification programs in place and are making honest efforts to meet the goals for declassification specified in the Executive Order. You're consulting with the user community for advice on declassification priorities. And you're cooperating with us on the review of records in our custody.

Mr. Carlin did, however, express the concern that the agencies within the IC should improve the way that they process other agencies' information for declassification. The executive order appears to have motivated significant action toward reducing the production of new classified information and accelerating the release of previously classified documents.

However, the implementation of the order is far from complete. Although Executive Order 12958 contains the goal of the declassification of most documents over 25 years old by the year 2000, some agencies have taken advantage of the exceptions to the rule. For instance, the CIA claimed that 64% of their classified information could qualify under the exemptions for automatic declassification, an argument that the DCI Historical Review Panel objected to in August 1996. The initial estimates of the amount of material that would require review for exemption tended to be rather high. Consider the estimate that the Air Force submitted to ISOO in 1995:

The total estimated amount of Air Force records which meet the criteria of Section 3.4 is: 70,598 cubic feet (176,495,000 Pages). Of the total figure, 70,288 Cubic Feet (175,720,000 Pages) contain potentially exemptible information. Specific record reviews over the next five (5) years will determine more precisely what records must remain classified and what records can be automatically declassified. The remaining 310 Cubic Feet

How these requests for exemptions are considered will have a large impact on the amount of material that is actually released under the automatic declassification program.

PROS AND CONS OF CLASSIFICATION IN THE POST-COLD WAR ERA

Is President Clinton's policy on classification sufficient for the post-Cold War era? Or did it go too far? In order to answer these questions, the advantages and disadvantages of classification must be considered in light of the fall of the Soviet Union.

The Cost of Classification vs. The Cost of Declassification

The Cost of Classification

Safeguarding classified information is an expensive task. According to Representative David E. Skaggs and ISOO, the cost of protecting classified information was about $5.6 billion in 1995. This estimate includes funds paid by industry for security involved with Government contracts, which consists of about half of the total. This estimate does not include the protection of classified information within the CIA, a figure that is itself classified. The indirect costs of classification are difficult to quantify, but a 1994 study commissioned by the Director of Central Intelligence and the Secretary of Defense suggested that they are much more than the direct costs of security:

The cost of security can be depicted as an iceberg having four facets. Two of the facets are visible and therefore more or less quantifiable. The other two are hidden below the waterline and, while difficult to measure, experience suggests they may be very large indeed. . . . the visible facets of the iceberg are made up of direct and indirect security costs. Together they account for a small percent of the iceberg. . . . Below the waterline are difficult to quantify and comparatively large hidden costs, loosely defined as inefficiency and opportunity costs. The commission believes that attacking these kinds of costs can yield near-term savings without degrading effectiveness.

The study also reported that the direct cost increases with the level of classification. Security for unclassified programs was typically one half to one percent of the total operating cost. In classified programs whose existence was acknowledged, this percent age was typically one to three percent. The security cost for unacknowledged programs varied widely, from one percent to ten percent, with one report as high as 40 percent.

Outside the Government, indirect costs of classification can also be found. In industry, the classification system imposes requirements in addition to the development of the requested product, which increases the cost to the Government. In testimony before the Commission on Protecting and Reducing Government Secrecy in Congress, an engineer from the "Skunk Works", the Lockheed facility where the U-2 and the SR-71 were developed, lists some of the complications that arise as a result of the classification system:

Overclassifying technology inhibits information exchange between programs and leads to "reinventing the wheel."
Classifying contractual and financial data within a corporation, which in today's environment should rarely be classified, inhibits accurate forecasting, limits oversight, and could eventually lead to an erosion in shareholder value based on unavailability of information for analysis.
From a legal standpoint, classifying unnecessary paperwork can put the company and the customer in jeopardy of union actions and lawsuits. . . .
For larger programs, massive declassification would not only eliminate unneeded protection for older information but would also have the potential for decreasing the amount of classified areas that are dedicated to storage.

Withholding Government sponsored research from the private sector prevents its use to improve the competitiveness of American industry. Michael Ravnitsky, technical director of the Industrial Fabrics Association International, contends that classified research could be valuable for the textile industry: "Decades of work done by the Defense Department and its contractors in the area of safety and protective fabrics would be of enormous use to our industry." Other technologies, such as microwave electronics, have benefited from the decision of the U.S. Government to reveal its research in those fields. The true cost of classification must take into account these negative impacts upon American industry. In academia, unnecessarily classified information prevents historians and political scientists from obtaining all of the information that they need to make a complete analysis. Trudy Peterson, acting archivist of the National Archive in 1995, would have preferred that some of the oldest of these documents be released with little review:

In an interview before the executive order was released, Peterson sug gested adding a provision that would force the release of information 50 years old or older, with no exceptions. "The President signed an executive order last November in which we declassified the majority of the materials that we hold in-house that predated World War II," she says. "And I don't believe that we have had severe repercussions in this government because we finally released all of the documents remaining that I know of from World War I. The Republic didn't fall."

Classified documents of historical value may impede academic research with little benefit to national security. One historian who gained access to some of the archives of the CIA expressed the opinion that the release of the documents there would not reveal any previously unknown episodes in the history of the IC:

Whatever remains in the CIA files cannot be nearly as awful as the Ameri can public imagines. To be sure, I hardly saw everything there was to see, but I got not even a whiff of dirty tricks that had somehow remained hidden from Church Committee investigators or the army of historians and authors who write about the CIA. I really believe that it would be in the Agency's interest to let historians see for themselves what remains classi fied. I do not see why the Agency does not declassify almost any secret that is more than 30 years old.

The declassification of historically valuable information might improve the quality of historical analysis. It would definitely reduce the need for historians to request the declassification of the information through the Freedom of Information Act, which may, in turn, free the staff who review FOIA requests to respond to more applicants. According to Anthony H. Passarella, the Director of the Directorate for Freedom of Information and Security Review within the Department of Defense:

A great deal of DoD time is spent processing FOIA requests for a rela tively small group of prolific requesters who continually seek vast quanti ties of classified, historical data, which has long been retired either to the Federal Records Storage Area in Suitland, Maryland, or been accessioned by the NARA. Retrieving records from the Records Storage Area takes time, and review of large volumes of classified material also takes time as mentioned earlier.

If documents such as these could be released automatically without detriment to national security, both those who request documents through FOIA and those who review them would benefit.

There is a danger that classification can be used to hide fiscal irresponsibility, environmental damage, and other embarrassing Government actions. Executive Order 12958 specifically prohibits such practices, but many recent examples suggest that the threat continues to exist. When former Government workers sued for allegedly unsafe working conditions at an Air Force facility at Groom Lake, the federal Government declared that many of their lawyer's documents were in fact classified, an action consid ered by several individuals to be a tactic to avert the lawsuit. Environmental damage caused by the early nuclear weapons effort remained classified for many years, as did human radiation experimentation, by the predecessors to the Department of Energy. Much information about these actions has since been released by the DOE as a result of Energy Secretary O'Leary's "Openness Initiative" that includes a directory of declassified DOE information on the World Wide Web.

Without sufficient information, public advocacy groups, lawyers, and concerned citizens do not have the opportunity to criticize the actions of the Government. This hurts the individuals affected, the Government, and the nation in general. Terry Anderson expressed his concern over the difficulty of retrieving information about his own captivity in Lebanon:

In such delicate areas, the Government has a right to national secrets. In my case, I was the focus of highly sensitive investigations. I am fully aware that some of the information I am seeking names intelligence agents and sources who cannot, and should not, be revealed. Where lives and national security are at stake, I am quick to justify caution. At the same time, I can't believe that much of the information in those files poses any danger to national security. The hostage crisis is over. It is time now to judge our leaders on how well they have done the jobs we elected them to do. We cannot judge them without information. Could it be that they do not wish to be judged, and that is why the system makes it so hard to come by that information?

Classification limits public oversight, leaving the entire task of oversight to the Legislative and Executive Branches.

The Cost of Declassification

The cost of the declassification process, however, must also be considered. In order to determine whether information can be safely declassified or not, the information must be reviewed by knowledgeable personnel within all of the Government agencies that may be affected by its release. Some estimates for the cost of this review process have been as high as $2 a page, although these have been attacked by the Information Security Oversight Office. For example, the Navy suggested in a memo to ISOO that "Calcu lated at $1.00 per page, the cost for declassification review could exceed an estimated $500 million." A much lower estimate is provided by the experience of the National Archive, which had declassified 16 million pages in three months in 1995 at a cost of about 3 cents per page. One of the reasons behind the discrepancy among agencies' estimates may be a difference of opinion toward the realism of "true automatic declassifi cation," the release of information without page-by-page review. Steven Aftergood, the Director of the Project on Government Secrecy at the Federation of the American Scientists, describes how both the high and low estimates are derived:

. . . the whole point of automatic declassification is that it is supposed to minimize the amount of resources that are consumed. To the extent that it genuinely is automatic and quantities of records are declassified in bulk, those records do not have to be examined page by page, they do not have to be handled by qualified reviewers, and so on. The resource problem comes because most agencies are reluctant to practice automatic declassification; they are insisting on exempting large fractions of their files and doing detailed review of those files - and that is extremely resource intensive.

The debate over the cost-effectiveness of the automatic declassification program in terms of a direct price for the Intelligence Community is difficult to resolve without detailed information about the nature of the classified documents eligible for release.

Another potential cost of declassification is the inadvertent release of information vital to national security. The case for declassification cannot be plausibly made if it depends on an infallible system. As above, however, knowledge of the specific subject matter of the classified information which may released would be necessary to make an informed estimate of this cost. One recent example of a possible error in declassification occurred on the Department of Defense's "Gulflink" earlier this year. Gulflink is a World Wide Web page that provides access to declassified information pertaining to the Gulf War Syndrome. In February 1996, several documents on the computer system were removed at the request of the CIA:

Mr. Tuite, who was the chief investigator in a Senate Banking Committee investigation of gulf war illnesses from 1993 until last year, said the report [suggesting U.S. Government knowledge of the exposure of American soldiers to chemical weapons during the Gulf War] was put on the Penta gon site on the Internet -- http://www.dtic.dla.mil/gulflink/ -- last July along with hundreds of other declassified documents. But the documents disappeared from the site in February. Pentagon officials said that the documents were removed from the Internet because of a dispute between the Defense Department and the CIA over whether the material had been declassified too quickly and whether they disclosed intelligence methods.

The documents were later reposted on the Internet, but they were also distributed unofficially by individuals who had downloaded the files during the period in which they were first publicly available, demonstrating the irreversibility of the flow of information in declassification projects. The danger of the release of national security information must be considered in any declassification scheme.

Developments in the Post-Cold War Era

Government secrecy has no place in a perfect world. However, neither do intelligence agencies; the existence of both is justified by the need to maintain national security in a far from perfect world. For a document to be classified, Executive Order 12958 requires that:

The original classification authority determine that the unauthorized disclosure of the information reasonably could be expected to result in damage to the national security and the original classification authority is able to classify or describe the damage.

The purpose of this section is to evaluate the damage that the release of classified information can have in the post-Cold War era. Some information can almost always be expected to have a sufficiently adverse effect on the security of the United States to justify its classification. For instance, the Government would seem to be almost always justified in maintaining the secrecy of knowledge that enables the nation's adversaries to develop weapons of mass destruction. In the post-Cold War era there are few absolute rules, however. For example, modern technology has enabled certain actions to be announced in advance with little detriment to the effort, as the Clinton Administration recently demonstrated with the announcement of the decision to launch cruise missiles against Iraq before the operation had begun. In every case, both the cost of classification and the risk of declassification must be considered and balanced.

New Requirements for National Security

Coalition Warfare

The increase in the number of multilateral military operations has caused new problems for the protection of information used on the battlefield. In the post-Cold War era, U.S. Armed Forces have cooperated with the militaries of several other nations and with the United Nations. In the process, the U.S. has had to share intelligence information with governments that would not otherwise be considered close allies. In contrast, during the Cold War the U.S. often operated with long-time allies such as Great Britain. Information provided to such nations could be assured to have similar protection as the U.S. intelligence community would have provided, because these countries had the same interests as the United States did. In recent U.S. military operations, the Armed Forces had to coordinate with nations such as Saudi Arabia and Syria and organizations such as the United Nations, where the bond of common interest and common security practices is weak, increasing the risk for the release of classified information. For example, American intelligence information was recently discovered unguarded in Somalia after it had been shared with the United Nations during U.S. operations there. DCI John Deutch de scribed the new challenges to intelligence producers in a memo concerning DCID 1/7:

We must recognize that military commanders not only require timely information at the collateral level but they often must share information with coalition partners who may not normally receive U.S. intelligence. In this equation it is not our military forces that have changed, but rather the nature of conflicts in which our military forces are involved. The world has changed and our provision of intelligence must change with it. We need an updated intelligence product line that is releasable from the inception--a flow of reporting based on all available sources that is written from the outset to be given broad U.S. dissemination and appropriate foreign release. The challenge is to extract meaning from all intelligence sources and package it for our various customers without betraying the very sources that will provide the next critical piece of intelligence.

The directive that Deutch issued, DCID 1/7, mandates that intelligence be produced in forms that are more easily declassified to the level at which it could be immediately used in low-security environments. In particular, DCID 1/7 recommends the use of a "tear line," a place in a document after which a less sensitive version of the information included in the document can be found. Such a tear line clearly delineates the extent to which information within the document may be shared.

Ethnic Conflicts and Human Rights Violations

Another development since the Cold War has been the rise in the number of ethnic conflicts. Strife in regions such as Somalia and Bosnia has captured the attention of both the conscience and the military of the United States. The impact of the human rights violations there on national security is difficult to quantify. However, Lt. Gen. James R. Clapper Jr., USAF, the Director of the Defense Intelligence Agency, alluded to this threat in a statement to the Senate Armed Services Committee in 1995:

Whether these kinds of conflict impact on U.S. interests is not for me to say. But as a purely factual matter, their numbers are increasing and at a minimum they will confront the world with humanitarian disasters involv ing millions of people. Thus far, these conflicts have had a relatively indirect impact on the West's "interests" -- largely directed toward our conscience and our urge to make things better. However, it will only be a matter of time before the impact, whether it is major refugee movement or some other phenomenon directed against one of our close allies, is much more direct.

U.S. intelligence assets can be useful in detecting human rights violations during such conflicts. However, the low priority that these conflicts necessarily receive within the IC when U.S. troops are not involved sometimes leaves American resources untapped. For instance, in the case of the mass grave at the Bosnian town of Srebenica in 1995, the Intelligence Community failed to make timely use of the information that they had gathered, as reported in the New York Times:

On July 13, as the press reported refugees' accounts of mass killings, a U.S. spy satellite passing over Bosnia recorded pictures of two fields in which hundreds of prisoners were guarded by gunmen. One of these was the stony field of the "rabbit hunt," where the 17-year-old Avdic was shot.

But no one saw that picture for three weeks.

Techno-thriller readers think spy satellites make the CIA all-seeing and all-knowing. In fact, U.S. intelligence collects far more data than it can analyze. One former National Security Agency director said gleaning hard facts from the avalanche of information was like trying to take a drink of water from a fire hose.

To the IC's credit, the pictures were eventually discovered and presented to the United Nations Security Council and the public, but human rights activists stated that had the images been revealed sooner they might have been useful in their investigation of the incident. Had this intelligence been available in some form to a wider audience within the IC and even outside of the Government, this it might have been identified in a more timely manner.

Arms Control

A number of international arms control treaties and agreements have emerged after the Cold War, and the verification of these treaties has increased the need to share classified information. One of the most important subjects requiring verification is the disarmament of nuclear weapons and the stewardship of the nuclear stockpile of the former Soviet Union. According to Joan B. Rohlfing, the Director of the Office of Nonproliferation and National Security in the Department of Energy, declassification has been important for these agreements:

Our declassification program has also played an instrumental role in furthering U.S. nonproliferation goals, especially in the initiative to achieve a bilateral agreement with Russia for the exchange of classified information to facilitate confirmation of transparent and irreversible nuclear weapons reductions.

The option to exchange classified information can make arms control treaty verification a realistic goal, something that might strengthen national security more than the protection of the shared information. Professor Frank von Hippel of Princeton University explains that:

It would be easier for the U.S. and Russia to collaborate on improving the security of fissile material if its management were more "transparent," that is, if each side declared its inventories of nuclear material and allowed the other to make spot checks of these declarations. Increased transparency would also make it possible to verify nuclear disarmament. To date, the verification of nuclear arms reductions has been limited to missiles, their launchers and bombers. It has not yet been extended to warheads and warhead components.

The need to prevent the proliferation of nuclear weapons and to reduce the stockpile of the former Soviet Union may outweigh the risk of sharing information with a former adversary.

Commercial Imagery

As the requirements of national security have changed, so have the threats to Government secrecy. One new threat, less sinister but no less dangerous than other threats, is the increased public availability of information around the world. In no area of intelligence gathering is this more apparent than in overhead satellite imagery. By the end of this year Earth Watch, Inc. expects to have launched a satellite with a resolution of 3 meters, and by the end of 1997 Space Imaging plans to have launched a satellite with a resolution of 1 meter. Over 20 satellites are planned to be launched by the end of the decade. By comparison, the best resolution for modern satellite imagery commercially available today is 10 meters, found in the French private satellite system called SPOT. Several commercial and public uses are anticipated for the sharper images, such as assistance in disaster response, urban planning, and crop analysis.

However, one potential use for the information is for independent verification of Government information. With commercial imagery, it would no longer be necessary for the media or public interest groups to accept the U.S. Government's information as the last word. For example, when Iraq invaded Kuwait in 1991, Peter Zimmerman used five- meter resolution images from a Russian satellite to substantiate the claim that the size of the invasion force had been smaller than the size reported by the U.S.. He stated that with one-meter resolution imagery, "I would have been able to make conclusions with extremely high confidence. I would have been able to see individual vehicles on the road." The ability of any government to keep information gathered by satellite imagery secret will be seriously impaired by these commercial initiatives. As commercial imagery becomes a reality, the need to maintain a high level of security for U.S. satellite images that have value for environmental research and the monitoring of activities that the IC considers a low priority may diminish.

RECOMMENDATIONS

Executive Order 12958 and Automatic Declassification

Executive Order 12958 made several significant improvements to the classifica tion system. It is too early to make a definitive judgement concerning the changes in the requirements for classifying information, but initial data suggest that the executive order is having the effect of reducing new classified materials. The program of automatic declassification currently underway has suffered from large requests for exemptions, but, again, the progress towards the declassification of historically relevant material being made through systematic review reported by ISOO implies that the very existence of an automatic declassification policy has motivated intelligence agencies to review their archives with renewed vigor.

The question of how the IC could support the distribution of large amounts of declassified material is an important one for automatic declassification. One argument to repeal the provision could be that the IC lacks the resources to distribute the information to the public. However, distribution options that have been developed for the FOIA system provide hope for answering this question in a cost-effective way. The National Security Archive at George Washington University, for instance, is a privately funded, non-profit organization that maintains a large collection of documents released through FOIA. The Archive maintains a public reading room and a database of FOIA requests, and it compiles collections of materials on popular topics on CD-ROM. The IC should seek advice from organizations such as these in order to develop an archive for automatically declassified information that would perhaps be only partially funded by the Government.

The large number of exemption requests that have been filed do threaten to slow the declassification process. However, all legitimate concerns for the need to protect certain information for longer than 25 years should be seriously considered; instead of limiting the number of requests that can be made, the Information Security Oversight Office should be expanded so that the requests can be evaluated in a timely fashion. If the goal of declassifying all appropriate information over 25 years old proves unrealis tic due to the exemption requests and the effect of the program is only to initiate a more time-intensive systematic review process within the IC, then the policy could still be considered a success, even though far fewer documents would be declassified. This approach is intended to minimize the cost to the Government and to the public of classified information while avoiding the inadvertent release of information whose secrecy is still required for national security. The progress made so far in declassification should not be squandered by requiring that automatic declassification meet its goal by the year 2000 over the objections of the IC. Rather, these objections should be carefully considered by an adequately funded ISOO.

The staff of ISOO is tasked to oversee the classification system within both the Government and industry. They also provide support to ISCAP, the body that will review applications for exemptions to automatic classification after the deadline for submitting exemption requests to the President has passed. ISOO is well-positioned to assist the President in evaluating the exemption requests, except for its small size. The release of their 1995 Report to the President in November 1996 suggests that they have inadequate resources to execute their tasks. According to Steven Aftergood, "They [ISOO] need to be increased in size by a factor of ten, although even double would be good. They have an annual budget of less than $1.5 million, so it is not unthinkable that it should be doubled, and I think that would go a long way towards achieving what the executive order sets out to do."

Perhaps the additional funding for ISOO can be found by requiring intelligence agencies who have requested that a large portion of their documents over 25 years old be considered for exception to automatic declassification to lend financial support to the office. A threshold could be established, and agencies that apply for more exemptions than the threshold amount would have to contribute resources to ISOO. This would have the additional benefit of establishing an incentive to reduce the volume of requests.

Reconsidering the Level of Classification of Government Information

Providing Information to Non-Traditional Consumers

DCI Directive 1/7, mentioned above, recommends that intelligence be produced in a way that enables the military to share it with non-traditional allies if necessary:

Coalition warfare is the future model for U.S. military conflicts, and I want the Intelligence Community to rededicate itself to the concept of releasable tailored intelligence -- intelligence produced at the lowest security level commensurate with the protection of sources and methods or produced in a format that allows for timely disclosure to U.S. customers or authorized foreign governments. From this point forward we must "write for the consumer."

DCID 1/7, which focuses on producing intelligence in a format that is easily releasable to foreign governments and international organizations, should be extended to explicitly include intelligence in the areas that the conference believes further cooperation with organizations outside of the IC would be beneficial. These areas may include environmental intelligence, health intelligence, support for law enforcement, and efforts to combat global organized crime. Although DCID 1/7 applies to all subjects of intelligence, the motivation for the directive, as conveyed in the memo accompanying it from DCI Deutch, was to improve support to military operations. The non-military subjects listed above should receive equal attention in the next DCI directive on this topic. In particular, the directive should include techniques analogous to tear lines for satellite imagery, such as reducing the resolution of the images, that would allow their distribution to a wider audience without revealing the capabilities of the sources of those images. Also, the DCI should recommend that ISPAC, a council of individuals outside of the Government who set priorities for declassification under E. O. 12958, include representatives from the fields listed above. In this way, the benefits expected for military operations under the directive can also be realized for non-traditional, but nevertheless important, intelligence objectives. Information could be shared more easily, resulting in the more effective utilization of intelligence gathered by the IC. Also, the initial preparation of intelligence with these consumers in mind would minimize the chance that information vital to national security would be inadvertently released.

Performance Measure: Unauthorized Leaks

One of the best measures of the shortcomings of the current classification system is the existence of unauthorized leaks within the Government. The unauthorized release of classified information indicates that individuals within the IC believe that some documents are improperly classified. Leaks pose a greater danger than systematic declassification, because it is more likely that a leaker would unwittingly release information that would pose a truly grave danger to national security. If declassification were faster and more reliable, or if certain documents were not classified in the first place, then it would remove an incentive to resort to leaks. The intelligence agencies should perform a thorough analysis of leaks to determine not only their sources but also the categories of information that are most likely to be released. No specific information should be considered for declassification simply because it has been leaked to the public, but a review of the classification policy for any category of information that is continually leaked may lead to an improvement of the system and the prevention of further leaks. Also, as the executive order and other reforms of the classification system are implemented, the frequency of leaks should decrease.

The recommendations made here and the recent changes in the classification system under Executive Order 12958 would help to restore the balance between the need for secrecy to protect the security of the United States and the need for the various consumers of U.S. Intelligence to use the information provided to them in the interest of national security. As new threats to national security emerge and as the Government gains expertise in automatic declassification, further changes may need to be considered. In any case, all of the members of the U.S. Intelligence Community should continue to keep as broad an audience as possible in mind as they attempt to learn the "truth" and inform others so as to "keep the nation free."

Aftergood, Steven. Interview. Director, Project on Government Secrecy, Federation of American Scientists, Washington, D.C., Nov. 5, 1996.

Aftergood, Steven. Secrecy and Government Bulletin. Issue 52, Sept. 1995.

Aftergood, Steven. Secrecy and Government Bulletin. Issue 55, Jan. 1996.

Aftergood, Steven. Secrecy and Government Bulletin. Issue 62, Oct. 1996.

Anderson, Terry. "My Paper Prison." The New York Times Magazine. April 4, 1993.

Brown, M. F.. Assistant for Information and Personnel Security, Office of the Chief of Naval Operations, Department of the Navy. "Memorandum for Director, Information Security Oversight Office." Nov. 30, 1995.

Cain, Sarah L.. "Eyes in the sky: satellite imaging blasts off." Photonics Spectra. Oct., 1995, Vol. 29, No. 10, p. 90.

Carlin, John W.. Archivist of the United States. "Keynote Address." Declassification Conference, Voice of America Auditorium, United States Information Agency, Washington, D.C., Dec. 5, 1996.

Carrier, Jim. "High-flying firm offers cameras in the sky." The Denver Post. Sept. 22, 1996, p. 8.

Clapper Jr., Lt. Gen. James R..,USAF. Director, Defense Intelligence Agency. Prepared Statement for the Senate Armed Services Committee, Jan. 17, 1995.

Clinton, William J.. Classified National Security Information. Executive Order 12958, Apr. 17, 1995.

Davidson, William A.. Administrative Assistant to the Office of the Secretary of the Air Force. "Memorandum for Director, Information Security Oversight Office." Oct. 6, 1995.

DCI Historical Review Panel. Report to the Director of Central Intelligence. Oct. 7, 1996.

Deutch, John. Director of Central Intelligence. "Memorandum on Producing Timely Tailored Finished Intelligence: Writing for the Consumer." Apr. 16, 1996.

Deutch, John. Director of Central Intelligence. DCI Directive 1/7. June 15, 1996.

Dupont, Daniel G. and Lardner, Richard. "The Culture of Secrecy." Government Executive. June 1995.

Engelberg, Stephen. and Weiner, Tim. "Srebrenica." New York Times. Oct. 29, 1995.

Garfinkel, Steven. Comments During Third Session of the DoD Historical Records Declassification Advisory Panel. Aug. 9, 1996.

Garfinkel, Steven. Interview. Director, Information Security Oversight Office, Nov. 5, 1996.

Gertz, Bill. "Deutch orders intelligence to be 'releasable' ; CIA chief seeks less secrecy to ease information swap with military, foreign nations." The Washington Times, May 1, 1996, part A, p. A8.

Hoberman, Henry S.. "Cold War secrecy should be put on ice." The Quill. vol. 81, no. 6, July 1993.

Information Security Oversight Office. 1995 Report to the President.

Joint Security Commission. Redefining Security: A Report to the Secretary of Defense and the Director of Central Intelligence, Washington, D.C., Feb. 28, 1994.

Lockheed Martin Skunk Works. Response to Commission on Protecting and Reducing Government Secrecy Request for Information, Sept. 13, 1995.

The National Security Archive. "The National Security Archive Home Page." http://www.seas.gwu.edu/nsarchive/. The Gelman Library, George Washington University.

Office of the Special Assistant for Gulf War Illnesses. "Gulflink: Gulf War Illnesses Home Page." http://www.dtic.mil/gulflink/.

"Openness Press Conference: Fact Sheets." U.S. Department of Energy. Office of the Press Secretary, June 27, 1994.

Passarella, Anthony H.. Director, Directorate for Freedom of Information and Security Review, Office of the Assistant Secretary of Defense (Public Affairs). Statement before the Subcommittee on Government Management, Information, and Technology Committee on Government Reform and Oversight, June 13, 1996.

Ravnitsky, Michael. Qtd. in Dupont, Daniel G. and Lardner, Richard. "Needles in a Cold War Haystack." Scientific American. Nov. 1996, p. 41.

Reagan, Ronald. National Security Information. Executive Order 12356. Apr. 2, 1982.

Rohlfing, Joan B.. Director, Office of Nonproliferation and National Security, Department of Energy. Statement before the Senate Appropriations Committee, Energy and Water Development Subcommittee, Apr. 16, 1996.

Shenon, Philip. "CIA Orders Inquiry Into Charges of Chemical Arms Cover-Up." New York Times. Nov. 2, 1996, Sec. 1, p. 9.

Shenon, Philip. "Report Shows U.S. Was Told in 1991 of Chemical Arms." New York Times. August 28, 1996, Sec. A, p. 1.

Stix, Gary. "Public Eye." Scientific American. August 1996, p. 18-20.

Thomas, Evan "A Singular Opportunity: Gaining Access to CIA's Records." Studies in Intelligence, Vol. 39, No. 5, 1996.

Von Hippel, Frank. "Fissile Material Security in the Post-Cold-War World." Physics Today. June 1995.

Weiner, Tim. "Lawmaker Tells of High Cost of Keeping Secret Data Secret." New York Times. June 28, 1996, sec. A, p. 20.

Wittes, Benjamin. "Suddenly, Your Briefcase is Classified." Legal Times. June 26, 1995.

Zimmerman, Peter. Qtd. in Gary Stix, "Public Eye," Scientific American, Aug. 1996.