WASHINGTON >> The Department of Homeland Security, at the urging of Congress, is building tools to more aggressively examine the social media accounts of all visa applicants and those seeking asylum or refugee status in the United States for possible ties to terrorist organizations.
Posts on Twitter, Facebook and other social media can reveal a wealth of information that can be used to identify potential terrorists, but experts say the department faces an array of technical, logistical and language barriers in trying to analyze the millions of records generated every day.
After the December mass shooting in San Bernardino, Calif., “we saw that our efforts are not as robust as they need to be,” Francis X. Taylor, undersecretary for intelligence and analysis, the top counterterrorism official at the Department of Homeland Security, said at a congressional hearing.
Travel industry officials and immigration rights advocates say the new policy carries the peril of making someone who posts legitimate criticism of U.S. foreign policy or who has friends or followers who express sympathy toward terrorists subject to unwarranted scrutiny.
Their concerns underline the mounting challenge for law enforcement agencies that are trying to keep pace with the speed and scope of technology as terrorists turn to social media as an essential tool.
“We haven’t seen the policy, but it is a concern considering the already lengthy and opaque process that refugees have to go through,” said Melanie Nezer of HIAS, a group that helps resettle refugees in the United States. “It could keep out people who are not a threat.” Several trade organizations in the travel industry said they had similar concerns.
The attackers in San Bernardino, Tashfeen Malik and her husband, Syed Rizwan Farook, had exchanged private online messages discussing their commitment to jihad and martyrdom, law enforcement officials said. But they did not post any public messages about their plans on Facebook or other social media platforms, officials said.
Since the shooting, counterterrorism officials and lawmakers have grown increasingly worried about the use of social media by terrorist groups like the Islamic State.
In Congress, several bills, including one by Sen. John McCain, R-Ariz., have been introduced that would require the Department of Homeland Security to screen the social media use of refugees and those visiting or migrating to the United States.
McCain’s bill, along with a similar measure introduced by Rep. Vern Buchanan, R-Fla., would require the department to examine all public records, including “Internet sites and social media profiles,” to determine whether an applicant would be a security risk.
“This legislation adds an important and necessary layer of screening that will go a long way in properly vetting the online activities of those wishing to enter the United States,” said Rep. Michael McCaul, R-Texas, chairman of the House Homeland Security Committee. “A simple check of social media accounts of foreign travelers and visa applicants will help ensure that those who have participated in, pledged allegiance to or communicated with terrorist organizations cannot enter the United States.”
Congress has yet to act on the legislation, but Democrats have also called for more screening of social media accounts. In December, 22 Democratic lawmakers sent a letter to the secretary of Homeland Security, Jeh Johnson, urging him to have the agency immediately begin reviewing applicants’ social media accounts.
“We believe these checks, focused on possible connections to terrorist activity, should be incorporated into DHS’ vetting process for visa determinations, and that this policy should be implemented as soon as possible,” the letter says.
The department said that it had developed a list of nearly three dozen situations where social media can be examined to screen applicants, but that it does not consistently examine accounts for signs of support or participation in terrorist organizations.
Last year, the department established a task force to study the issue and to make recommendations to Johnson on ways to incorporate the screening of social media accounts into the vetting process. A new social media policy is expected soon, although agency officials did not give a timetable.
The department says it has four pilot projects to examine the use of social media among applicants for immigration benefits. One of the projects, which began in December and runs through June, screens the social media accounts of applicants for the so-called fiance visa, the program under which Malik entered the United States.
The U.S. Citizenship and Immigration Services, which is part of the Department of Homeland Security, uses social media as part of the screening process for Syrian refuges, but only when the person is flagged because of a hit in an intelligence database or when questions are raised during an interview with immigration officials.
The agency does not have the ability to reach private communications sent via direct messaging.
“All we have is the ability to access the public-facing statements that individuals make,” Leon Rodriguez, the agency’s director, said at the congressional hearing, on Feb. 3. “We do not have a way to reach private communications.”
Rodriguez said that while some social media platforms like Facebook and Twitter were scanned with software, most screening was done manually by analysts at the agency.
“It’s slow going and, short term, is just for Syrians,” he said.
In the future, the Department of Homeland Security hopes to find a way to fully automate going through a huge amount of messages and other data, he said. It said it would bring in experts from the technology industry at the end of the month to talk about the tools available to search social media.
The agency said it would also have to hire and train additional people, including more linguists, to screen the accounts.
Experts say that analyzing the social media accounts of tens of millions of people who enter the country each year to work, visit and live would be exceedingly difficult.
Users might have multiple accounts on different platforms, the accounts can be in different languages and many people use aliases.
John Elder, a data-mining expert based in Charlottesville, Va., who has worked with the IRS and the Postal Service on fraud detection, said it was possible to build models to scan thousands of social media postings for information that could help identify potential terrorists.
“But hopefully people won’t have the expectation that it will be an infallible oracle,” he said. “It will help you pinpoint things you need to check further, but it’s not error free.”
David Heyman, a former assistant secretary of Homeland Security for policy, agrees. “You have to be careful how you design the proposal to screen people,” he said. “Artificial intelligence and algorithms have a poor ability to discern sarcasm or parody.”