To Keep Top Secrets Secret, We Need Fewer of Them
Fewer secrets shared with fewer officials are essential to keep our secrets secret.
Late last week, Senate Intelligence Committee chairman Mark Warner described how over-classification of national security-related information is a key and neglected factor behind the latest, disturbing intelligence leak. “We need frankly a system that limits classification to really important documents and then have a process to declassify when appropriate.” His argument: fewer secrets shared with fewer officials are essential to keep our secrets secret.
He’s on to something. Just before the New York Times first revealed that top secret documents had been posted on social media, my nonprofit released a major study on over-classification. Those in the “know” understand that over-classification is bad but insist officials can’t help themselves: the penalties for letting a document leak far outweigh any professional rewards that might come from making secret information more available.
That’s the conventional wisdom. It’s also dead wrong. In fact, effective national security organizations have strong incentives not to over-classify. There are effective ways to avoid doing so, and one of the most important U.S. intelligence agencies—the National Geospatial-Intelligence Agency (NGA)—has done it.
This is the key conclusion of a classification project my nonprofit ran for two years. Last month, the Public Interest Declassification Board (PIDB) asked that I privately brief them on the project’s final report. The study subsequently received national attention. It gave examples of how harmful over-classification has become:
American troops in Afghanistan couldn’t get timely, properly classified imagery from our government. To fight effectively, they had to buy unclassified imagery from private firms, which they could share with their Afghan compatriots.
The Pentagon recently used a new information restriction—controlled unclassified information (CUI)—to keep unclassified weapons test results from Congress.
The head of the U.S. Space Force was prohibited from publicly uttering the name of our key spy satellite system (KH-11), even though the media has long and repeatedly referred to it.
Official historians writing classified studies to help black program managers learn from past mistakes and successes have been unable to find key classified documents because routine archival classification reviews, which would otherwise assure proper document filing, are not being done.
Wasteful, expensive, duplicative military space programs have been shielded from oversight by special access program classification barriers that effectively block cross-communication and information sharing.
Draconian security rules discourage innovative firms here and abroad from sharing their best with the Pentagon. These same rules are blocking needed military collaboration with our closest allies.
Previously unclassified U.S. civilian nuclear export and cooperation information is being kept from Congress even though such commerce might help countries develop nuclear weapons options.
What allows this? In a word, mismanagement.
Currently, our government has over 2,000 security classification guidebooks and roughly 1,400 original classification authorities. Nobody can consult them all, and they don’t. These numbers, and the impossibility of mastering their guidance, are why so many government officials over-classify—it’s safe, it’s easy, and it takes little or no thought.
Fortunately, our best government organizations disagree. A case in point is the NGA. Seven years ago, it recognized it could only succeed if it added value to its imagery and got it to its “customers” quicker than commercial, unclassified space imagery firms. If it continued to deliver its product too late, or made it difficult to share with critical allies and firms, the agency would effectively go out of business.
The NGA was floundering using sixty-five classification guidebooks to classify its imagery. So what did it do? It boiled these down to a single electronic guide, eliminating previously subjective, contradictory guidance. It also required classifiers to justify their proposals to an intra-agency group of users, de-classifiers, historians, and subject-matter experts, and made appeals easy and quick. Finally, it encouraged constant updating of its new consolidated guidebook.
One Senator—Mike Rounds (R-SD)—noticed. He asked the Pentagon to report on how well the NGA’s example is being followed. He saw its model as the one our government should replicate. If it doesn’t, automating the review of the millions of classified documents it generates will be pointless: even the best document filtering system will fail if it follows contradictory and vague guidance that’s inevitable with thousands of different guidebooks and officials acting as classifying authorities.
So, what’s yet to be done? The Biden administration has pledged to tighten up the current system. But Congress must also act. At a minimum, it should fund and authorize dedicated staff to its own declassification policy review unit, the PIDB, which it has so far failed to do. Congress almost did this last year. It needs to get on with it.
Second, Congress should task the PIDB to oversee any government bidding on advanced technology contracts aimed to help automate the classification process. It also should track how many guidebooks, original classifications guidebooks, original classification authorities, classified documents, and declassification requests are being generated and recommend how best to reduce these numbers.
More, of course, is needed. But skipping these first steps will only ensure we still will have too many classified documents to track to ever keep America’s top secrets from leaking again.
Henry Sokolski, executive director of the Nonproliferation Policy Education Center, served as deputy for nonproliferation in the Defense Department and is the author of Underestimated: Our Not So Peaceful Nuclear Future.