One Social-Media ‘Like’ Doesn’t Say Much About a Potential Extremist
Data from open-source social media is generally only helpful in combination with other information.
One social-media “like” may be grounds for punishment under the Pentagon’s new anti-extremism rules—but it’s far from certain that a single click would trigger an investigation or, really, say much useful at all.
On Monday, Defense Department officials outlined several lines of effort that the Defense Department took to get at the problem of extremism in the ranks, including re-examining the department’s insider threat program and screening procedures for recruits. Pentagon spokesman John Kirby said that continuous monitoring of social media “was not the intent” of the new policy and that the government didn’t have the ability to monitor “private” social media.
However, the government can monitor public social media just as any advertiser or member of the public can. Indeed, the Defense Counterintelligence and Security Agency is charged with continuous vetting across the military, including civilians. But searching millions of accounts is not an effective way to find an extremist, according to a former senior Defense Department official who worked insider threat detection for the military.
“You’re talking about millions of people here. At volume, it just doesn’t work very well,” said the official.
Nor is a single post on a social media site a good indicator of an actual insider threat.
About three years ago, the Defense Department experimented with using open-source social media posts to find troops who might be committing crimes, the official said.
“We did some tests on a small percentage of cases. We ran some open source research to get some leads,” the official said. “We shipped it out to the field investigators that were working the case and to say that it was almost of no value may be giving it too much credit. It wasn’t helping at all.”
Still, social media posts can help answer questions about someone who has already displayed other problematic behaviors, such as financial difficulty, trips to unusual places, trying to access information not related to job function, an arrest record, etc. The best place to find good indicator data is in reports from people close to the individual, said the former official.
“It’s going to be in the records that says this person has been accused by his coworkers of yelling things at them that are inappropriate at work. Or he’s maxing out his company credit card, buying inappropriate things. His performance sucks. It’s a lot of little things that say ‘this guy is going off the deep end.’”
The new policy makes clear that some social media behaviors, such as liking some posts that make threats against the government or lawmakers, coud be met with punishment.
“Liking is an advocation,” said Kirby.
What punishment would that merit?
“It’s going to be very case-specific and it’s going to be up to the chain of command,” he said. “They know their units…better than anyone.”
But punishment is likely to take the form of a conversation or intervention, not a dismissal, according to the former official, precisely because it’s only one small piece of a larger puzzle.