Miami-Dade County

Here’s a way to track would-be terrorists before an attack

A growing memorial of flowers line the grass in the Seneff Arts Plaza in downtown Orlando, to honor those killed at Pulse nightclub in Orlando.
A growing memorial of flowers line the grass in the Seneff Arts Plaza in downtown Orlando, to honor those killed at Pulse nightclub in Orlando.

After the mass shooting at an Orlando nightclub earlier this month, questions have abounded about how it could have been prevented.

The discussion has raised a familiar topic: the logistical and legal questions surrounding social media monitoring. Researchers at the University of Miami say their recently released research can help track would-be terrorists before they strike, although the protection of peoples' civil liberties remains an ongoing debate.

The UM researchers say their findings, published June 17 in the journal Science, can help focus law enforcement resources on those who are highest risk. The study, called “New online ecology of adversarial aggregates: ISIS and beyond,” applies a mathematical equation that describes the movement of particles — whether the dance of dust specks or the ebb and flow of bird flocks — to the world of online social media activity.

The idea is to track how single online users form communities around a shared interest, how these groups evolve over time, and who really drives the action.

“Lots of people chatter about ISIS,” said physicist Neil Johnson. “There's a huge amount of noise — and none of them have any influence.”

One of the team's main findings was that informal online communities often form and then grow in a recognizable pattern that, when plotted, resembles a shark fin.

“We saw that the lead-up several months before a sudden, unexpected [event] is characterized by an explosion in the rate of creation of aggregates” ad hoc online groups, he said. He stressed that even though such groups may be small, what matters most is the speed at which they suddenly start growing. This rapid change in growth rate produces the curve of the “fin,” while the sharp drop-off represents when such groups are shut down by tech companies or law enforcement agencies. The information can thus help authorities to figure out which groups to zoom in on.

Perhaps most relevant are the study's findings about so-called lone wolves, or people who may be inspired to commit an attack from their consumption of online content without receiving any material or logistical support from groups. These types of individuals have been an acute source of anxiety for some in the wake of the Orlando shooting.

The Miami researchers found that such individuals are never truly alone for any extended amount of time.

“Any one of these members could look like they are alone in a snapshot [in time],” he said. “But they were — and they will be again — in an aggregate,” or online group. A better understanding of how such individuals move in and out of groups can help authorities track a person and see what types of conversations they are privy to.

When asked about the civil liberties concerns often raised in response to government tracking online, Johnson said that the research helps address such fears.

“I think it protects them,” he said. “This [research] removes the people in these groups that are just ranting occasionally.”

He said that as an academic, he had not spoken directly with law enforcement, but he hoped his work would be helpful. His research on this topic originated from his previous efforts to track political unrest in Latin America, and could be applied to track any type of group-forming online. He stressed all of the data his team used was open-sourced, such as public Twitter accounts.

However, not everyone is convinced such tools can be used appropriately.

Hugh Handeyside, a staff attorney at the ACLU's National Security Project, said that while he was not familiar enough with the UM study to comment on it specifically, he was generally concerned about automated algorithms used to monitor online speech.

“Identifying speech requires context and intent,” he said. He added that while authorities have long used automated procedures to identify and shut down child pornography — which is by definition illegal — such processes could not extend to content that is more subjective. “There is nothing wrong in itself with having radical or extreme views,” he noted.

Meanwhile, some lawmakers in Washington have also been quick to respond to the shooting from a national security standpoint with a focus on online activity. House Foreign Affairs Committee chairman Ed Royce, R-Calif., on Monday announced a hearing for this week entitled “Countering the Virtual Caliphate.” The committee includes Florida lawmakers Rep. Ileana Ros-Lehtinen, R-Miami, Rep. Ron DeSantis, R-Fla., Rep. Ted Yoho, R-Fla., and Rep. Curt Clawson, R-Fla.

“Through the Internet and social media, ISIS is inspiring attacks within the United States and projecting force across the Middle East and Europe,” read a statement from Royce. “Yet this administration’s response continues to be disjointed and inadequate.”

However, an aide to a Democratic committee member who spoke on condition of anonymity insisted the hearing was not political.

“It's absolutely a policy-based hearing,” he said, calling the committee's efforts “bipartisan.”

He said the hearing would focus on efforts overseas due to the committee's jurisdiction, and would examine ways to “empower credible voices that give a counter-narrative” in online spaces where radicalized groups operate. These tactics could range from official State Department tweets aimed at young people intrigued by ISIS to YouTube contests to amplify anti-extremist voices from communities where radicalized groups recruit to “fight the false narrative about Islam” that such groups propagate.

The effort, he said, would ideally involve not only the State Department but also tech companies like Twitter and Facebook, which could lend their expertise on how to track the reach of a given message among a target audience.

But Handeyside flagged a potential issue, saying that ACLU had filed public records requests due to the organization's concerns that the federal government is pressuring such tech companies to delete the content that the government itself is not allowed to delete.

“To the extent [the government] is doing looks like an end-run around the First Amendment,” he said.

Seamus Hughes, deputy director for the Program on Extremism at George Washington University's Center for Cyber and Homeland Security, who will testify at the hearing, said such partnerships were key. Because free speech concerns “makes government hesitant to get into that space,” he said, the Obama administration has thus far pursued a strategy of training American civil society members — religious leaders, non-profit organizations, etc — to do the counter-messaging against radical groups. He added a plug for the Department of Justice to write up some legal guidelines to address concerns from local leaders who want to be engaged in such efforts but are “worried about ending up on some list.”

Such guidelines currently do not exist. “It's the wild west,” he said.

For Handeyside, that point circled back to a larger problem.

“That just points out the problem with monitoring,” he laughed. “It chills protected speech. People are afraid they will be flagged for engaging in the conversation.”

If the federal government were to issue such a directive from the DOJ, he said, “that would be a sad commentary on the efforts they're engaged in.”

Note: The national and Florida-based FBI office declined to answer specific questions for this story, although the national office did refer for more information to its Domestic Investigations and Operations Guide. The section on protection of civil liberties and privacy highlights several guiding “principles,” including investigating for a specific purpose, and not using race, religion, ethnicity or national origin as the sole basis for launching an investigation — although such factors “may be taken into account under certain circumstances.”