By Spencer Ackerman
Wired.com
The military is scrambling to identify disgruntled or radicalized troops who pose a threat to themselves or their buddies. So the futurists at Darpa are asking for algorithms to find and pre-empt anyone planning the next Fort Hood massacre, WikiLeaks document dump or suicide-in-uniform.
This counterintelligence-heavy effort isn’t Darpa’s typical push to create flying Humvees or brainwave-powered prosthetic limbs. But the Pentagon’s far-out R&D team has made other moves recently to hunt down threats from within.
The idea behind the Anomaly Detection at Multiple Scales, or Adams, effort is to sift through “massive data sets” to find the warning signs of looming homicide, suicide or other destructive behavior. “The focus is on malevolent insiders that started out as ‘good guys.’ The specific goal of Adams is to detect anomalous behaviors before or shortly after they turn,” the agency writes in its program announcement.
Currently, Darpa says, the Defense Department doesn’t actually know how “a soldier in good mental health” actually comes to pose an “insider threat,” defined as “an already trusted person in a secure environment with access to sensitive information and information systems and sources.” (WikiLeaks, anyone?)
“When we look through the evidence after the fact, we often find a trail –- sometimes even an ‘obvious’ one,” Darpa adds. “The question is can we pick up the trail before the fact, giving us time to intervene and prevent an incident? Why is that so hard?”
Adams is supposed to fill the breach. But what kind of tech would be necessary to detect these anomalies? What sort of data actually represent worrisome anomalies, as opposed to a soldier harmlessly venting steam?
Unclear. The full Adams request for proposal won’t be released until “mid-October.” It’s inviting anyone interested in thinking those questions through to an Oct. 19 “Industry Day” conference at the Arlington, Virginia, offices of the Systems Planning Corporation, a defense-research firm.
Adams isn’t Darpa’s first move into internal military sleuthing. In August, it announced plans to build a system called Cyber Insider Threat, or Cinder, to hunt down patterns of suspicious cyberbehavior on military networks that might indicate internal subversion or outside infiltration.
But even though Cinder is overseen by top hacker Peiter “Mudge” Zatko, it doesn’t sound like Darpa knows precisely what it’s looking for. Anomalous behavior could be “comprised of entirely ‘legitimate’ activities, observables and the data sources they will be derived from,” its August contractor solicitation acknowledged. How to distinguish the real warning signs from the false positives?
Then again, the rest of the Defense Department has been caught off-guard on radicalization. After a shooter — allegedly Army Major Nidal Malik Hasan — killed 13 people at Fort Hood last November, an independent inquiry ripped the department for not giving clear guidance on how to identify the warning signs of radicalization.
And that inquiry came under fire from the House Armed Services Committee for avoiding questions about radical Islam taking hold within the military after it came out that Hasan was in touch with al-Qaeda-aligned preacher Anwar al-Awlaqi.
All this suggests the blind are still leading the blind when it comes to stopping internal military subversion. It’s far from clear what kind of data — troops’ e-mail? web trails? book orders? — Darpa would use to ferret out troops who pose a risk to themselves or others.
Nor is it clear if any such effort can succeed against a soldier who just snaps. But it’s not as if there are clear alternatives to confront an insidious and deadly problem.
View Original Article With Links HERE
No comments:
Post a Comment