Artificial Intelligence (AI) and Machine Learning (ML) have seen a recent uptick in focus within the Department of Defense. As attention and funding of AI and ML initiatives grows, so will the business opportunities. According to IDC, AI/ML spending is expected to reach $57.6 billion globally by 2021, up from the $12 billion in 2017. Future potential is top of mind in DOD and the plans have been drafted to establish a center dedicated to providing AI and ML solutions across multiple agencies.
DoD agencies are primarily focused on using AI and ML to crunch data, help make more informed decisions and combat potential near-peer adversaries. DARPA, the research arm of the DoD, announced on September 7 AI Next, a $2 billion initiative aimed at substantially improving AI research and adoption rates within the department. DARPA has historically supported AI technological advancements dating back to the early 1960’s and is currently funding 20 AI-related programs.
AI Next is the second major DoD AI initiative announced in the last few months. In June, the Pentagon announced the Joint Artificial Intelligence Center (JAIC). The center will have oversight of all service and defense agency AI initiatives. The JAIC will report to CIO Dana Deasy, but all projects under $15 million will remain under the authority of the hosting agency.
Deasy has been instructed to provide a list of initial “National Mission Initiatives” for JAIC to focus on by September 27, 2018. The first of those AI initiatives to be addressed is expected to be Project Maven. Google recently announced it would be withdrawing from the Maven contract after employees at Google vehemently opposed the used of Google AI to analyze surveillance video that could lead to the application of lethal drone strikes. However, expect to see the DoD replicate Maven for non-tactical applications.
According to DARPA, the DoD will improve AI capabilities by focusing on the following areas:
New Capabilities: AI solutions are routinely leveraged to enable DARPA research and development projects, with more than sixty benefitting in 2018 alone. Specific areas that will benefit from AI and MI applications:
- Real-time analysis of cyber-attacks
- Detection of fraudulent imagery
- Construction of dynamic kill-chains for all-domain warfare
- Supply chain analysis
- Complex data mining
- Biomedical research
Adversarial AI: Machine learning systems are currently susceptible to cyberattacks. Data supplied to train these systems is the biggest weakness for AI as of now. The DoD and industry must address these issues at scale as more AI solutions are deployed and relied upon.
High Performance AI: Increased power and power efficiency are recurring issues faced with increased AI/ML application. DARPA is currently researching AI-specific hardware designs and best practices to reduce requirements for “labeled” data training.
Next Generation AI: DARPA believes that pioneering the next generation of AI/ML algorithms that allow computers to explain their actions, solve problems alongside humans and reason with “common sense knowledge” are of paramount importance for maintaining AI dominance.
In addition to AI Next, DARPA announced the Artificial Intelligence Exploration (AIE) program, a strategy built around quick turnaround of high-risk, high reward projects. The goal is to leverage streamlined contracting and funding mechanisms to establish start dates within three months of opportunity announcements. Additionally, DARPA is hoping to help researchers deliver new AI solutions within 18 months of the award. With AIE, DARPA will focus on “third wave” applications of AI, with the focus of creating machines that understand context and adapt to ever changing environments.
With all the focus on AI and ML, there will be plenty of opportunity for vendors with related capabilities to participate, but they need to stay on top of what’s going on because DOD is moving fast.
Want to hear more about AI trends in the federal government? Join us at the 5th Annual Government IT Sales Summit, November 15, 2018 in Reston, Virginia.
By Ryan Granato
immixGroup, An Arrow Company
Last modified: May 3, 2019