Posted February 14, 2013
By Scott R. Flick
At this stage in the media cycle, few could have missed the news of several Michigan and Montana TV stations airing an EAS alert warning the public of a zombie attack. As I noted earlier this week, while the facts surrounding these alerts are still developing, it appears they were the result of someone outside the U.S. triggering the stations' EAS equipment via that equipment's Internet connection. While the resulting burst of media stories quickly devolved into a flurry of zombie jokes, the movie that came to mind as the story developed was not Night of the Living Dead, but the Terminator films, which feature an interconnected national defense network called Skynet. In the films, Skynet becomes so sophisticated as to turn on its creators, causing a nuclear launch that brings destruction to the human race and, after the movie, Arnold Schwarzenegger to the California Governor's Office.
For many years, the EAS system, as well as its predecessor, the Emergency Broadcast System, operated by having a number of primary broadcast stations connected to governmental agencies through a closed network (typically over telephone lines). When an alert was sent to these primary stations, they would broadcast the alert, which would then be picked up and aired by stations monitoring the signal of the primary station, and in turn, by other stations monitoring those secondary stations. This created a daisy chain in which an announcement over one station quickly spread to stations throughout the alert area.
One of the perceived flaws of the Emergency Broadcast System was the amount of human interaction it required. For example, when a national alert was accidentally triggered in 1971, it caused little disruption, since many station managers intercepted it and did not air it because they heard no corroboration of the emergency over their newswires. While it turned out that those station managers were correct in concluding it was an accidental alert, critics of the Emergency Broadcast System counted this event as a failure of the system, since the delay inherent in station managers deciding whether an alert should be aired (and the risk that they may reach the wrong conclusion) puts more lives in danger.
The shift to EAS from the Emergency Broadcast System was done largely to increase the automation, and therefore the reliability, of the system. That digital squeal you hear accompanying an EAS warning is a digital code instructing other equipment, including the public's radios (if properly equipped), to activate, lessening the chance that emergency alerts go unheard, either because a link in the daisy chain failed to relay the message, or because the public was not listening to radio or watching TV at the time.
The downside to this level of automation soon became apparent. As I wrote in September of 2010, a radio ad for gas stations sought to satirize emergency alert announcements, right down to including the EAS digital tone. Because EAS equipment has a poor sense of humor and is no judge of context, any station airing the ad would trigger EAS alerts on the stations "downstream" from it in the EAS daisy chain. For this reason, Section §11.45 of the FCC's Rules provides that "No person may transmit or cause to transmit the EAS codes or Attention Signal, or a recording or simulation thereof, in any circumstance other than in an actual National, State or Local Area emergency or authorized test of the EAS." Just a few months later, the problem repeated itself when TV ads for the disaster movie Skyline included an EAS tone among the many sound effects in the ad.
The highly automated nature of EAS was demonstrated yet again this week, when a Wisconsin radio station's morning show disc jockeys played a tape of the zombie EAS alert, including the digital tone. The result was--you guessed it--the alert being automatically rebroadcast over at least one local television station whose EAS equipment was activated by the digital EAS tone.
While the automatic nature of EAS creates the risk of false alerts propagating rapidly, at least the false alerts up until now were somewhat self-inflicted wounds, caused by either the system being erroneously activated by a governmental mistake, or by an EAS Participant accidentally airing an activation code contained in third-party content. Because of the closed nature of the system, false activations necessarily required a mistake from a participant in the EAS system, even if that mistake was airing third party content that had not been screened for EAS tones.
This week's episode, however, appears to have been something entirely different. In an effort to expand the types of consumer devices capable of relaying an alert, the backbone of the EAS system was moved not long ago from the closed network model to an Internet-based system. The benefit is that mobile and other devices connected to the Internet will be able to relay alerts to the public automatically, ensuring the broadest possible distribution of the alert. The bad news, however, is that by shifting to an Internet backbone, we have opened the public alert system to the same outside forces that plague every other aspect of the Internet. In this week's case, it appears that someone outside the U.S. spent a number of days trying to use those Internet connections to access station EAS equipment. In at least a few cases, they succeeded, generating the now-infamous zombie alerts.
So the good news is that we are well along in the development of an automated emergency alert system that can spread emergency information to most Americans in a matter of minutes. The bad news is that by putting the system almost entirely under the control of "the machines" (a Terminator term), the moderating effect of human involvement is greatly limited. In addition, by connecting this equipment through the Internet, we have expanded the ubiquity of the system, but at the cost of making every EAS Participant's equipment, whether in Michigan, Montana, or elsewhere, readily accessible to every miscreant in the world with an Internet connection.
Thus, we are perfecting an automated response system that operates most efficiently without human involvement, while creating opportunities for control of that system (or at least portions of it) to fall into the hands of those who do not have our best interests at heart. In other words, Skynet is now a reality. This Skynet does not, thankfully, have the power to initiate nuclear launches, but it certainly does have the capability to launch public panic. A more realistic alert than a zombie attack could cause immense confusion and harm, particularly where the false message is being reinforced by identical EAS alerts on every source of information available, whether it be broadcast, cable, satellite, or smartphone.
I have worked with many of the individuals who created and have dedicated themselves to improving and expanding the current EAS system, and I have no doubt that they are moving quickly to seal off any vulnerabilities discovered in the zombie attacks. Still, I can't help but wonder if EAS is now subject to the same Internet arms race that bedevils online security everywhere, with ever-evolving measures and countermeasures being deployed in an effort to stay one step ahead of those wishing to commandeer the alert system for their own benefit or amusement. If so, the questions becomes: which is worse, false alerts that panic the populace, or a populace that becomes so used to false alerts that they ignore a real one?