Are American spies the next victims of the Internet age?
- By Daniel B. PrietoDaniel B. Prieto is a member of the Aspen Homeland Security Group, a senior associate for homeland security and counterterrorism at CSIS, the innovation fellow at the Center for National Policy, and a former executive for a Fortune 50 technology company.
The furor over the NSA’s data collection and surveillance programs has been fierce, but the debate it has sparked has been far too narrow. The value and legality of the programs is extremely important, but we should be asking a more fundamental question: Does the U.S. intelligence apparatus, transformed in the dozen years since 9/11, meet the challenges and threats that the United States faces today and into the future? Is the "business model" of U.S. intelligence — how intelligence is gathered, analyzed, and used, basically the key operating features that comprise the U.S. intelligence enterprise — sufficient and sustainable, or does it need to evolve to something different or something more?
The 9/11 attacks and the wars in Iraq and Afghanistan provided new missions, a new focus, and ready laboratories for testing, building, and honing counterterrorism and counterinsurgency techniques. At the same time, the global rise of Internet, mobile, and social technologies — along with the Big Data that those technologies generate — allowed the intelligence community (IC) to innovate, implement, and sharpen new techniques for tactical intelligence and targeting. Many of the targeted drone strikes in Afghanistan, Pakistan, and Yemen, for example, relied on social network analysis that included cell phone call data in combination with human intelligence. And link analysis of cell phone records and behavioral analysis of social media and Internet use are key components of the PRISM program.
But for all the IC’s tactical innovation in support of military or covert operations, it has yet to make similar transformative strides to leverage new technologies and new operating models to generate strategic intelligence. How might the IC use the same Internet, mobile, and social technologies to revolutionize how it develops better foresight into things like the Arab Spring, the Syrian civil war, or the recent civil unrest that has rocked Turkey?
The opportunity for dramatic transformation cannot be overstated. The traditional operating model for intelligence production is a lot like the "old media" operating models in the private sector. In those models (think: print books or music CDs), intellectual capital was produced, owned, and tightly controlled by a few actors in hierarchical organizations, and the physical product was disseminated through fixed distribution channels. For the IC, information gathering, analysis, and the creation of strategic judgments occurred in a similar manner. Controlled intellectual capital came in the form of stolen secrets. Intelligence was produced and vetted by a small number of professionals, and their distribution of the final product tightly controlled.
The hard times that have befallen newspapers, record labels, and book publishers serve as a cautionary tale for what lies ahead for U.S. intelligence if we fail to modernize not just technology, but the whole "business model" of how intelligence is produced in the 21st century. Wikipedia killed Encyclopedia Britannica. Craigslist and eBay killed newspaper classifieds. iTunes is killing the music CD business. And Kindles and iPads are killing print books. If the intel community clings to the old, closed, secrets-based model for developing strategic intelligence, the cost won’t be in dollars and cents, but in our security as a nation.
To be sure, the IC has taken some steps to revise its operating model. After 9/11, it established fusion centers to move information out of silos and encourage collaboration between different agencies. Fusing the analytic and operational capabilities of multiple agencies in one facility, with unified leadership, provided improved capabilities to field commanders.
At the same time, fusion centers are not and cannot be the evolutionary apex of IC collaboration. Information integration needs to be about much more than putting analysts in the same room and asking them to work better together. Fusion centers do not take full advantage of the power of networks, which characterize the information age. They are like the clunky online chat rooms of the late 1990s, when today we need the speed, flexibility, and crowd-sourcing power exemplified by tools like Facebook and Twitter.
In addition to fusion centers, the U.S. intelligence community in 2005 created the DNI Open Source Center to increase the availability of open source information for the intel community. While this was a positive step, it was a long time coming. There had been repeated calls, starting at the end of the Cold War in the late 1980s and through the 1990s, from military and congressional leaders, to create an open source intel center. More recently, the DNI has begun to invest in studies of how to use big data to improve strategic intelligence and foresight.
Despite these innovations and investments, the IC still operates behind a shroud of secrecy and relies on ostensible secrets. Strategic intelligence is still largely created the old fashioned way: in a closed system, with limited inputs, long production cycles, and limited distribution. In short, the business of strategic intelligence in 2013 is still more similar to than different from book and newspaper publishing and music production in 1999.
To remake strategic intelligence for the threats and volatility we face today, the U.S. intel community needs to shift its relative focus on controlled, sanctioned, and vetted secret information toward a greater focus on the knowledge that is available from all sources, especially from an open world. While this is getting done in pockets, they are still separate entities and activities. Open source intelligence is still outside the core intelligence process, a bolt-on to a bygone operating model. Open source is viewed as valuable, but still a novelty act, an adjunct to classified intelligence, and it is not fully integrated into the creation of strategic intelligence.
Learning about what’s happening in Egypt today from diplomatic cables might be useful, but developing systematic ways to mine social media from Cairo and encouraging analysts to engage commentators, analysts, NGOs, bloggers, and academics who are not government employees, might allow the IC to better sense trends from the Arab Spring and street attitudes from Tehran to Tashkent. It will require a significant change in the culture, attitudes, and current practices of the intelligence community, but in a world where less knowledge is secret and more secrets are public, the benefits are compelling.
Obviously, the 24/7 news cycle, RSS feeds, and Twitter users can provide and catalog news more quickly and in greater volume than the clipping services provided by the Foreign Broadcast Information Service during the Cold War. But beyond speed and volume, platforms like Twitter offer a new "social intelligence" that captures the intentions, actions, and sentiments of non-government actors in real time. The window these technologies provided into the 2009 Iranian protests and Green Movement, the Arab Spring uprisings, and the 2013 protests in Turkey all demonstrate their power.
The value of social media to the intelligence process may not be immediately apparent — particularly for senior intelligence leaders who did not grow up with these technologies and who may not use them even now. So, it is worth speculating about how valuable these technologies would have been during the Cold War. For example, how much better could our intelligence have been if Twitter feeds and mobile downloads of YouTube videos had been available during the 1968 Prague Spring political uprisings in Czechoslovakia? Or if intelligence analysts had an opportunity, over years, to observe or engage academics, on the ground NGOs, trade union activists, and Catholic religious leaders through social media during the rise of the anti-Soviet Solidarity movement in Poland during the 1980s? Similarly, how important might this kind of open source social intelligence have been in helping the United States predict the fall of the Soviet Union?
In a 2008 book on U.S. intelligence, author John Diamond argued that "the scorn heaped on the [Central Intelligence] Agency in the early 1990s … is based on the dubious assumption that predicting the breakup should have been an easy call." In reviewing the Diamond book, Georgetown University professor Roger George noted that while the CIA forecast the risks posed by Gorbachev’s attempts to reform Soviet socialism, it never "fully appreciated the centrifugal forces at play in Soviet society." He argues further that U.S. intelligence "accurately gauged many military Soviet programs, but undervalued the overall strain defense placed on the economy, and it identified the falling quality of life as a major threat to stability but never questioned Moscow’s ability to control the pressures."
The comparative advantage of traditional secret intelligence lies in areas of diplomatic and military activity, precisely the areas where the CIA analysis of the Soviet Union excelled. The weakest areas of analysis — regarding social strain and economic discontent — are exactly the areas where digital, social, and mobile technologies have their strong suit. In our Cold War thought experiment, the wisdom of the crowd could have made a difference to the work of American intelligence analysts. They would have been better off with a heavy infusion of open source intelligence and social media — both at arm’s length (e.g., just reading Twitter feeds) and more closely (e.g., engaging observers and analysts via social media).
Today, the use of technology, social intelligence, and analytics has allowed the United States to dramatically degrade al Qaeda through better targeting. But by removing (Egypt) or threatening to remove (Syria) a wide range of authoritarian governments from power, the revolutions of the Arab Spring have set free a wide range of Islamist political movements and actors. While many of these groups may satisfy themselves by focusing on local politics, it remains to be seen whether any of the nascent Islamist groups have an interest, like Hezbollah, in supporting attacks on the United States or its interests. Adding to the uncertainty is the influx of foreign fighters into Syria, who at some point will return home with significant battle experience, just like Osama bin Laden returned home after fighting the Soviets in Afghanistan in the 1980s. The range of countries and new political actors the United States must get knowledgeable on quickly is staggering, ranging from Mali in West Africa to Somalia in East Africa to Libya and Egypt in North Africa to Pakistan and Afghanistan in South Asia and back through the Gulf through the Middle East to the Levant.
One way to deal with the challenge is to take a traditional intel route, dramatically increasing human intelligence on the ground. This is already occurring with the Defense Intelligence Agency announcing plans in late 2012 to field "as many as 1,600 ‘collectors’ in positions around the world, an unprecedented total for an agency whose presence abroad numbered in the triple digits in recent years," as the Washington Post reported. For all the merits the move might possess, it will be costly and take significant time to recruit, train, place, and embed these operatives in the field. To be sure, technology is being used to augment these efforts: U.S. drone hubs now exist in Qatar and the United Arab Emirates to monitor the Persian Gulf, and in Djibouti, Ethiopia, Niger, and Seychelles to cover Africa. This surge of new human and aerial assets, while formidable, will excel at gathering traditional classified intelligence.
Better use of open source intel from digital, social, and mobile sources would dramatically speed and improve our ability to understand what is going on in these remote corners of the world. The opportunity is amplified by the extraordinarily rapid growth of mobile and Internet use globally. In late 2012, there were over 5 billion mobile phones in the world but only 1 billion smartphones. Within the next five years, there will be 5 billion smartphones. That rapid growth trend holds true in even in the most far flung places. At the time of the U.S. invasion after 9/11, Afghanistan had 500,000 cell phones; by late 2012, it had 18 million. Libya, of all places, is among the countries with the highest per capita cell phone use.
The strategic intelligence opportunity in these regions comes not simply from monitoring social media use, but also from a raft of analytics and open data innovation that, at the moment, is being driven by development organizations. The U.N. Global Pulse team has applied text analysis tools to Twitter, Facebook, blogs, online forums, and news sites to better forecast food crises in developing countries several weeks in advance. A former private-sector colleague of mine has done pro-bono work for the World Bank to analyze the intensity of light in open-source time-series nighttime satellite images to accurately forecast poverty levels. At a recent White House event hosted by the Obama administration’s chief technology officer, I participated in a working group of technologists and development professionals to brainstorm open source data indicators that could provide early warning of genocides.
The potential for change in the IC based on analyzing never-before available sources of open source data is enormous. It will allow an increase in data-driven judgment as analysts replace abstraction (strategic judgment in the face of scarce data) with enumeration (strategic judgment based on rich sources of data). The change will be similar to the revolutions spurred by the rise of analytics in professional sports, notably baseball and basketball. Instead of scouts focusing on a gut feeling of a prospect’s intangibles — think of a batter having a "pure swing" in baseball or of a power forward having an "NBA-ready body" in basketball — newly available data and statistical techniques, much of it innovated and curated by non-sports-professionals, allows sports executives, managers, and analysts to compare players’ "value over replacement player" or "alternate win scores." Sports knowledge and foresight has moved increasingly from abstraction to enumeration because of new open source data, the rise of non-traditional expert communities, and advances in the ability to analyze data for trends and previously unobserved connections.
Change will not be easy. And intelligence traditionalists may react negatively to an influx of open source data, the engagement of more outsiders, and more data-driven techniques — just like baseball traditionalists bridled against advanced baseball metrics, as chronicled in Michael Lewis’s book Moneyball. Similar changes in the IC will require significant, and likely difficult, cultural and behavioral change. Luckily, the biggest factor working in favor of this change is inexorable: the massive generational shift inside the intel community. More than half of American intelligence analysts joined the community after 9/11. Analysts in their early 40s bridge the analog and digital eras, having had mass market access to personal computers starting in high school. Analysts in their early 30s have had mass market access to the Internet and web for all of their adult and professional lives. Twenty-somethings are almost digital natives, having had access to digital technology since they were children and social and mobile technology since their teens. Over time, the adoption of new digital, social, mobile, and big data techniques will be second nature to the pool of U.S. analysts.
At the end of the day, the value of strategic intelligence to policymakers is foresight. Foresight does not mean prediction — CIA analysts armed with social media might still have missed the fall of the Soviet Union. But they would have had better insight into the social and economic conditions that precipitated it. When it comes to the Green Revolution in Iran or blowback from the Arab Spring, the social, mobile, and digital revolution, paired with "big data" analytics may not be a silver bullet to address our need for better strategic intelligence. But, given the uncertainty we face in so many unstable regions, it would be worrisome if U.S. strategic intelligence were left to confront 21st-century threats while still largely relying on a 20th-century business model.