Correction: A previous version of this story mentioned an individual by the name of Clark. This has been corrected to reference Cook, whose job title has also been corrected.

WASHINGTON — With the goal of providing military commanders and policymakers with the best possible analysis, defense intelligence has reached a point where innovations in information technology and cyber present an opportunity to drastically reimagine the entire enterprise, according to a Defense Intelligence Agency expert.

Due to the inherently complex environment, providing the right intelligence to support decision-making, processes and answering requirements is more challenging than its ever been, Louis Werdebach, senior defense intelligence expert for command, control, communications, computers and intelligence at DIA, said Wednesday at the C4ISRNET annual conference.

Over the past several years, DIA has been working "very hard trying to get ahead" of intelligence requirements gaps, "trying to close a number of the hard problems," Geoffrey Strayer, the chief of DIA's Office for Analytic Enterprise Operations in the Directorate for Analysis, said.

This has involved exploring new ways to create intelligence assessments, develop crowdsourcing, object-based production (a concept to allow everyone to consistently talk about particular intelligence), data science, artificial intelligence, and trying to figure out how to get ahead of the "huge" influx of information moving through the intelligence framework, Strayer said.

The sourcing of intel: How to meld data for a clear picture

Since 9/11, the government has recognized the importance of connecting the dots. Doing so to support command and control on the battlefield brings heightened need for accuracy and speed — a requirement the services struggle to manage.

An intelligence enterprise will need to move at a faster and faster pace, and analysts are already being given the technology to make decisions or assessments faster. What used to take an airman three hours can now take 15 minutes, Maj. Gen. Bradford "B.J." Shwedo, the commander of the 25th Air Force, said at the conference. Some of that is due to airmen taking the initiative to improve the technology they use, according to Shwedo.

But moving quickly still requires processing the intelligence coming in with a "suspicious eye," especially acknowledging the limitations of technology. "You have to constantly go, ‘OK, what am I missing here? I can’t trust that piece [of intelligence.' "

Steve Cook, the program manager for battle management, command and control at Northrop Grumman, is currently tasked to focus on innovative solutions to help optimize the massive defense intelligence enterprise for users like the 25th Air Force.

Cook said Northrop is focused on data analytics and finding actionable intelligence out of "big data" — massive amounts of information streaming in from countless sensors and sources.

From a soldier with a video camera all the way to satellite capability, "there is a lot of data," Cook said, and the challenge is sorting through all of it.

Other technologies that could aid the enterprise include:

  • Computer automation, where the computer can do more of the easy, shallow tasks and leave the deep thinking and decision-making to the intelligence analyst
  • Cognitive machine
  • Artificial intelligence
  • Autonomy

Man-machine teaming is another area that could be developed and implemented in the intelligence enterprise.

"What thinking can I turn over to the computer so I don’t have to," Cook asked, and how can the machines be controlled or commanded using such things as gestures like a hand movement.

There are few challenges related to incorporating innovative technology into the intelligence enterprise.

Trust is a big one, according to Cook. Users of such technology need to understand the gap in terms of human thinking and machine processing. For instance, Cook said, there’s the example of an Alexa, Amazon’s voice controlled system, ordering a dollhouse for a little girl because it overheard a conversation in the other room.

Overtrusting a system is another issue and can lead to fratricide when a human trusts the computer is thinking more than it can, Cook noted.

And there’s the issue of assuming nothing’s there if sensors aren’t picking anything up or transmitting intelligence to the right place. "How do we know what we don’t know," Cook asked.