Internet giants Facebook, Twitter and Google took center stage in Washington, D.C. on Wednesday at a congressional hearing aimed to move forward in solving the problem of foreign influence campaigns on American social media networks.
“We simply haven’t done enough,” Twitter co-founder and CEO Jack Dorsey said in a hearing focused on the twin failings of Silicon Valley and the U.S. federal government to deal with an intensifying global problem.
“We were too slow to spot this and too slow to act,” Facebook’s Sheryl Sandberg said. “That’s on us.”
Despite the regular infusion of mea culpas, Sandberg and Dorsey touted improvements by both companies in combating foreign influence including, most pointedly, the recent removal of hundreds of accounts across multiple independent foreign campaigns. Last week, Facebook banned Myanmar’s commander-in-chief as ethnic violence continues in that country.
Larry Page, the chief executive of Google parent company Alphabet, declined an invitation to attend the hearing. That left an open seat with a Google placard posted in front, while several senators condemned the company for failing to send a top executive. Google did offer to send senior vice president for global affairs and chief legal office Kent Walker in Page’s place.
“I’m deeply disappointed that Google, one of the most influential digital platforms in the world, chose not to send its own top corporate leadership to engage this committee,” Sen. Mark Warner, D-Va., said.
Both Dorsey and Sandberg emphasized the importance of artificial intelligence and machine learning in their nascent capabilities. While both companies say they are investing heavily in human resources like threat investigators, artificial intelligence and machine learning techniques are instrumental in almost every investigation carried out by the Silicon Valley firms.
“We’ve decided to focus a lot more on the behavioral patterns we’re seeing across the network,” Dorsey said. “While we can’t always recognize in real time where someone might be coming from or if they are representing someone who does not exist, we can see common patterns of behavior in utilizing the network to spread their information. We have been building a lot of machine learning and deep learning technology to recognize these patterns and shut them down before they spread too quickly. And also to link them to other accounts that demonstrate similar patterns. We’ve gotten a lot more leverage out of that in terms of scalability.”
Facebook has long touted, both publicly and privately, its use of artificial intelligence and other automation in combating rule-breaking content. Last year, Facebook’s Vice President of Global Policy Management Monika Bickert said that 99 percent of the “ISIS and Al Qaeda-related terror content” Facebook removes is detected before it’s ever flagged and in some cases before it’s even live, a result achieved “primarily through the use of automated systems like photo and video matching and text-based machine learning.”
In July, Facebook’s director of cybersecurity policy Nathaniel Gleicher described the use of automation “to do that shrinking of the haystack — to simplify the environment, to identify the patterns, and to run automated tools that can make the low-grade noise, the low-grade malicious activity harder to get through, and can make it easier for our sophisticated investigators to know where to look.”
“But you don’t just trust automated systems and artificial intelligence,” he said. “That’s why we have a threat intelligence team of incredibly well-trained investigators which focuses on the types of threat actors we see around the world and the types of activities we see. By combining those two, that’s how you get the best results from what we see.”
In ramping up defenses against foreign influence operations, U.S. officials have turned to allied governments. In a meeting last week, officials from the “Five Eyes” governments – Australia, Britain, Canada, New Zealand, and United States – agreed to collaborate more closely on the issue, according to Homeland Security Secretary Kirstjen Nielsen.
“More broadly, I have directed a shift from a ‘counterterrorism’ posture at [the Department of Homeland Security] to a much broader and wider ‘counter-threats’ posture to make sure that we’re doing everything possible to guard against nation-state interference,” Nielsen said Wednesday at George Washington University.
“We’re working on hashtag databases,” she said, “trying to identify IP addresses where we have continued to see nefarious activity emanating.”
Sean Lyngaas contributed reporting.