In today’s post-9/11 era, almost every U.S. government agency has been affected by the war on terror. For consumers, TSA security checks and the presence of the National Guard at major events are continuing reminders of the increased level of government attention and effort toward national security. As part of this shift, the government has had a more urgent need to access and understand content in a number of different languages to maintain effective global intelligence. This need has coincided with a major explosion of the use of new communication technologies like cell phones and Internet. Though not as visible to everyday people as the enhanced airport security checks, government translation has significantly evolved during the past decade, propelled by intelligence efforts and supported by new translation technologies.
Pre-9/11 Translation
For some history on the topic, prior to the war on terror, government translation was compartmentalized. There was an assumption that by carefully tracking influential individuals, agencies could predict and prevent most threats. The thinking of the time was still strongly influenced by the 20th-century combat model of warfare, involving infantry troops, tanks, strategic locations and the like.
Agencies targeted specific individuals for surveillance and translated only that information which they deemed most important. Independent language specialists were hired to translate conversations, documents and other information. Rather than distributing their knowledge into translation memory databases as they do today, translators physically wrote interpretations of words and phrases and then that passed onto others, who wrote up a summary of the translated content.
Globalization Already Underway
As late as the nineties various agencies within the government were still using this translation approach even as globalization was re-defining old paradigms. In the 1990s, the private sector adopted the mantra “Go Global.” Outsourcing and offshoring meant that businesses had partners, suppliers and customers all over the world. In order to communicate quickly and effectively, enterprises adopted a combination of machine and human translators. Through computer-assisted translation, computers were assisting humans in new and effective ways, using data resources like translation memory and terminology databases in order to speed up the work of translators.
Likewise, national security was being affected by a quickly evolving picture of terrorism. In his 2002 testimony, FBI official Dale L. Watson described “an apparent shift in operational intensity from traditional sources of terrorism -- state sponsors and formalized terrorist organizations -- to loosely affiliated extremists … paralleled by a general shift in tactics and methodologies among international terrorists that focus on producing mass casualties.” There was a paradigm shift taking place, which included the efforts of Osama bin Laden and the Taliban organization. Yet as a 2002 Subcommittee on Terrorism and Homeland Security report uncovered, at this time pre-9/11 agencies did not have the manpower and capabilities to translate all of the information that came their way to monitor these groups.
The September 11 Catalyst
After 9/11 hit, agencies extended their definitions of threats and the tactics used to combat them. Intelligence teams started to recognize the fact that seemingly unrelated events, for example the genocide in Darfur, might impact the safety and security of U.S. citizens at home and abroad. In other words, terrorists might be linked to active political organizations in other countries, or they might find willing recruits amidst economic ruin. All types of incoming intelligence and news became fair game for translation. Yet translating all information that might pose a threat to US security was a daunting task.
Information from around the globe needed to be efficiently translated so that agencies could make fast, informed decisions. Putting data into a queue to be translated by select individuals with institutional knowledge was far too time consuming and represented too much of a risk. The government’s new translation load, which covered all forms of information, from graffiti to cell phone conversations to email communications, overwhelmed the capabilities of individual human translators.
The Resurgence of Machine Translation
Government agencies had to scale up their translation capabilities, and in doing so they turned to the private sector. Enterprise machine translation was equipped with useful technologies that enabled quick translation in bulk. Humans would share information with machines in order to build up their translation memory, enabling computers to be more accurate.
These technologies were just what the government needed. With money specifically allocated to the war on terror, the public sector began to contract with the private sector. Companies like BBN, a think tank around language and language processing using computational models, became mainstays for ongoing translation. BBN, now Raytheon BBN, would monitor and translate continuous news broadcast by Al Jazeera.
By contracting with and purchasing technologies initiated in the private sector, the government was able to re-build its translation process into an always-on, efficient process that incorporated machines and humans. In this way, agencies have been able to quickly decide which data is actionable, rather than putting some information on the back burner based on prior assumptions. Today, technology is moving in the direction of machines growing smarter, with computers doing the bulk of the translating while humans work on higher-level, more nuanced material.
The Right Tools for the Job
Translation has gone from a function done by specialists to one that relies on the widespread use of computer intelligence and human-machine communications. Through this combination of machine and human translation, the government can now better address the modern terrorist threat. Machines are capable of translating in bulk the river of information that flows into the United States intelligence system every day. Humans work to perfect and modify machines’ translations and bring clarity to alerts that might arise. With this sequence in place, agencies can enable the right people to make actionable decisions on incoming information. Government terrorism specialist John Rollins has described modern terrorism as “a diffuse global network and philosophical movement composed of dispersed nodes with varying degrees of independence.“ With an enemy as nebulous and widely distributed as this, today’s combination of machine and human translation is a much more viable weapon for the job of intelligence.