The military has become extremely effective when it comes to generating data. In fact, you could argue that it has grown too effective at it.
With the emergence of today’s advanced sensors, and innovations in intelligence, surveillance, and reconnaissance (ISR) solutions across each domain – including land, air, sea, space and cyber – today’s military is capable of generating terabytes of information every single day.
Some of that data could be useful for the military – helping to identify potential threats to national security and mitigate threats to our warfighters. Some of that data could be relatively useless. It’s essential that the military is able to determine which data is important, and draw actionable intelligence from it.
However, according to a new Thought Piece by Jeff Kimmons, Ronald Makuta and Graham Gilmer of Booz Allen Hamilton, entitled, “Remaking Intelligence Processing Exploitation and Dissemination,” this incredible amount of available data is creating new challenges and problems for the military.
The sheer amount of data, the siloes making it impossible to analyze data from different sources or timeframes, and a lag in analyst recruitment are creating a situation where the military struggles to effectively process, exploit and disseminate (PED) the data that they’re generating in a timely fashion. It’s effectively equivalent to trying to drink from a firehose.
Unfortunately, the potential ramifications of not exploiting this data to generate actionable insights could be downright catastrophic. This is a problem that could literally be the difference between life and death for America’s warfighters, citizens, or allies.
However, Jeff, Ronald and Graham do highlight a potential solution. New technological advancements are available that can help the military overcome this PED challenge and truly take advantage of the data that they’re generating. These technologies include artificial intelligence (AI) and machine learning.
Bolstering a small staff with smart robots
With a lack of staff and a figurative mountain of data to sort through, it’s essential that human personnel only analyze and interpret the video, data, or information that may be relevant. New machine-learning and AI solutions are capable of handling the high level analytics – sorting and doing basic analysis of video and other data.
This leaves the actual analyzing of the data to military personnel and could eliminate the need for intelligence and military personnel to spend time watching and analyzing terabytes of video and data just to see if it’s relevant and important.
AI accomplishes this by giving machines the power and capacity to analyze video and data for red flags. These red flags can take the shape of unexpected movements or changes in the location of items, people or vehicles on video footage. Or, it could be the detection of a repeatable pattern in data or activity that illustrates something of relevance or interest. Once something is flagged by the AI, intelligence and military personnel can then be brought in to analyze it and determine if it is truly actionable intelligence.
Making AI a reality in the DoD and Intel communities
Ultimately, AI and machine learning solutions can effectively empower the defense and intelligence communities to do more with less by allowing their personnel to focus their time and attention on only the ISR data that is deemed important. But embracing these technologies across these communities may not be so simple.
The Thought Piece authors identify a few roadblocks and challenges that could make embracing AI and machine learning easier said than done. However, they did provide a handful of important considerations and steps that the DoD and Intel communities can take to make advanced AI a reality in their organizations.
First, to make AI a reality and effective, the military needs to break their data out of the siloes that currently exist. Right now, data is aggregated by disparate agencies and organizations and stored in disparate silos – independently owned and managed networks and data repositories where data can’t be easily shared and compared with the data from other organizations within the community.
Next, they need to reevaluate their existing PED technology and infrastructure and embrace new technologies. Much of the benefits that come from today’s advanced AI and machine learning solutions require the compute and storage power that can only come from the cloud. However, cloud migrations aside, there are other technology acquisitions and implementations that are required across these organizations that are imperative to these capabilities.
The final – and potentially most difficult – step that these organizations need to take involves a cultural change within their own organizations. They need to shift their organizational thinking to place higher value on situational understanding than situational awareness.
They also need to change the perception of AI and machine learning as an automator and job killer that will make personnel unnecessary or redundant. Instead, they need to position AI and machine learning as a way to make analysts better and more effective. Only then will they get the organizational buy-in necessary to make AI and machine learning a tool that can improve the efficiency of the entire enterprise.
When it comes to generating data, the defense and Intel communities have done an incredible job – too good of a job. But to truly gain actionable intelligence that can be used to identify threats, help drive military strategy, and improve operational effectiveness, they need help with the PED of that data. They don’t have the resources or the staff to do it alone – at least not without today’s advanced AI and machine learning solutions.
