USAF maturing AI-enabled counterintelligence, insider threat capability

A graphic from Cogility Software, describing its Threat Hierarchical Complex Event Processing streaming analytics platform, developed for the USAF, to detect potential insider threats and espionage. (Cogility Software)

The US Air Force (USAF) is working with industry partners to further mature an artificial intelligence (AI)- and machine learning (ML)-enabled counterintelligence (CI) software programme, designed to identify potential insider threats among service personnel with access to highly sensitive and classified information.

USAF officials with the air service's Concepts, Development, and Management (CDM) office agreed to a USD44 million development deal with California-based Cogility Software in September 2023 to develop the Threat Hierarchical Complex Event Processing programme.

The basis of the programme is to leverage Cogility's continuous intelligence platform software to collect and track “psychosocial and behavioural observations” and other associated data inputs from service and contractor personnel with access to USAF classified data, said Stuart Booth, chief operating officer at Cogility.

“It's a streaming analytics platform, and our special secret sauce is the analytics, or real-time analytics, and then we give the [human] analysts tools to do further drill down into investigative analysis to kind of come up with a result,” Booth explained to Janes.

“The differentiator is that [the continuous intelligence platform] can detect signatures that may not be found within typical AI applications. That's one of the reasons we are in the counter-insider threat programme, because human behaviour is very difficult to predict,” he added.

That collected psychosocial and behavioural data then goes through an automated analytics process, and the results are further reviewed by a human CI analyst, who can then determine the insider threat and where it is located, he said.