The head of the FBI’s Cyber Division said today that the bureau is making the “nation so much safer” by leveraging artificial intelligence (AI) and natural language processing technology in its tip reporting.
Cynthia Kaiser, the director of the FBI’s Cyber Division, said during GDIT EMERGE in D.C. on June 4 that the tip line will continue to employ humans, but emphasized “people miss things” that AI can catch.
“After it’s looked at by a human, after it’s flagged one way or another, we’re using natural language processing models to also go over the synopsis of the text of what that phone call or online tip then details to see did we miss something,” Kaiser said. “That AI is trained on the expertise of people who’ve been … taking in these tips for years and years and years and know what it is to flag.”
“That’s a way in which AI is making us so much safer and making the nation so much safer,” she continued, “There’s always a human in the loop with FBI when we’re doing that, but it helps us fill in the cracks.”
Kaiser said that there are “really basic” and “exciting ways” the FBI can leverage AI now – like information sharing and rapidly drafting text – that will save her team time on administrative duties.
“That’s a huge win for everybody,” she said. “There’s some really basic things that AI can do now that are really exciting to me, and it’s very labor saving to the folks that instead of having to write up something in a different format, they’re using their cyber skills to find adversaries.”
On the flip side of that, Kaiser also highlighted that AI is equally labor saving for cyber adversaries, like China, Russia, Iran, and North Korea.
She emphasized that AI is allowing them to become more sophisticated in their attacks, particularly when it comes to deepfakes and spear phishing campaigns.
She also said that AI is helping the nation’s adversaries become more obscure by evading detection and staying hidden. Kaiser said that with the rise of AI, she is seeing “a crop of adversaries who are becoming at least mildly better.”
The FBI’s cyber lead said the agency is looking internally to figure out how to evaluate not only how adversaries understand AI and are leveraging the technology but also what the actual threat level looks like.
“What we’re looking to do internally is to really identify what’s the framework for us to evaluate the actual threat,” Kaiser said. “Maybe this adversary integrated AI over here, [but] did it actually make them better?”
Kaiser said the FBI wants to be “able to really map, is China better today because they used AI? Is Russia better? How?”
The FBI cyber lead said knowing the adversaries’ actual level of threat is going to allow them to “complete the AI cycle” and better defend their networks.
“That’s really where AI right now shines because … that balance right now is weighted towards it’s better for the defense of our networks than it is a threat,” she said.
