Address by Lt General Dave Deptula (Retd), USAF Dean of the Mitchell Institute for Aerospace Studies, VA, USA, Guest of Honour at the e-sym on AI for Air Warriors
Well, good afternoon, it’s a real pleasure and a privilege to be with you today.
Let me start with a basic premise: militaries around the world are in the midst of a transition from the industrial age of warfare to the information age…from an age where we focused on individual weapon systems and inflicting attrition, to one where success depends foremost on the ability to manage one’s own information and decision-making while degrading that of the adversary.
If you accept this premise, then the potential revolutionary impact of artificial intelligence or AI on future operations should be readily apparent since AI, at its core, is an advanced method of information processing.
Indeed, a key takeaway from the recent U.S. National Security Commission on AI was that AI will be integrated into virtually all future technology in some form.
The US military approach to AI emphasizes human-machine teaming as a force multiplier for human operators and analysts—not to replace human operators with AI, regardless of what Elon Musk claims. The value added is where machines and people are not competitors; but instead, where humans are freed up to execute more highly dimensional tasks. Specifically, the next generation of AI-enabled systems will augment humans to deal with complexity, ambiguity, and speed.
With that in mind, there are some promising near-term applications for AI in air operations. I’ve got three general areas to highlight for you today. The first is automated operator assistance… particularly for intelligence, surveillance, and reconnaissance or ISR functions. One approach is to integrate AI to assist operators and analysts to sort through the vast amount of ISR data collected into actionable insights at machine speeds. This was the intent of Project Maven, which used AI to help analyze full-motion video data.
An alternate approach is to integrate AI on the front end of systems, requiring the operator to yield some decision-making to an AI-enabled virtual co-pilot. In this construct, the AI would reduce the cognitive burden placed on pilots by automating certain functions, enabling the human pilot to focus on battle management and other tasks better suited to human operators. Furthermore, by processing data at the battlespace edge, it also reduces the data bandwidth and storage requirements.
This construct was demonstrated for the first time in the US last winter, when the Air Force allowed an AI co-pilot to control sensor and navigation systems on a U-2 surveillance plane, with no pilot override.
Cognitive Electronic Warfare capabilities for tasks such as spectrum sharing and management, link adaptation, and electronic attack are another promising area. Applying AI in this realm provides the ability to execute optimal strategies and adapt to a changing environment at machine speed.
The second area is unmanned systems. These are likely to see greater AI-enabled autonomy moving up the “value chain” of what have traditionally been human tasking. For example, remotely piloted aircraft will unlikely be able to maintain secure communications with their ground control stations in contested environments. To address this limitation, AI-enabled UAVs can detect threats and other changes in the operating environment and then use their AI algorithms to decide on appropriate actions when cut off from their human operators.
Taking this further, the United States is actively working to develop loyal wingmen UAVs through initiatives such as the Air Force’s Skyborg Vanguard program and the Defense advanced research project agency’s Air Combat Evolution effort.
Teamed with manned aircraft, a human pilot can give general inputs and guidance to a swarm of UAVs and be confident that the swarm will behave as required. Absent AI-enabled autonomy, a single operator would not be able to manage that many platforms.
The emerging family of AI-enabled aircraft could fulfill a range of missions in support of advanced fighters and other manned aircraft. For example, they could act as an offboard sensor or provide additional ordinance capacity for manned aircraft. Autonomous aircraft could additionally conduct electronic warfare activities or serve as a communication gateways for the kind of combat cloud concepts of operation that are proliferating.
This last example points to a third natural application for AI: command and control. The reality is that legacy command and control or C2 systems currently in use were not designed for the speed and complexity that information-age all-domain operations demand. Furthermore, the evolution towards highly centralized C2 architectures that rely on robust, high-bandwidth communications will likely be disrupted against peer adversaries.
Artificial intelligence will feature prominently in efforts to modernize C2. AI-enabled decision support tools would allow commanders at diverse C2 nodes better control distributed forces; adapt to environmental or adversary actions; and impose complexity on enemy decision-making at speed. AI-enabled C2 would leverage the respective strengths of humans and machines; humans providing flexibility and applying their creative insights, and machines providing speed and scale while ensuring all relevant details are tracked.
I would also like to note the critical importance of partners with AI-enabled capabilities. As the US National Security Commission on AI pointed out, “uneven adoption of AI will threaten interoperability and the political cohesion and resiliency of U.S. partners.”
Given the problem-specific and data dependent nature of AI technologies, cooperation with partners and allies requires working through common objectives and common equipment. It will also be critical to accelerate agreements on architectures and standards, share technical expertise, and pursue AI use-cases for exercises and wargames.
At the same time, peer competitors are investing heavily in artificial intelligence. For example, China seeks to be a world leader in AI, with the aim if making the industry worth $150 billion. According to China’s 2019 Defense White Paper: “The form of war is accelerating toward an informationized warfare evolution, there are indications intelligentized warfare is emerging”
The Chinese Peoples Liberation Army objective is to use AI algorithms, machine learning, human-machine teaming, and autonomous systems collaboratively to paralyze its adversaries. The Chinese seek to use AI to deliver precise effects to immobilize their adversary while defending their own system-of-systems.
Russia is similarly pursuing AI-enabled capabilities, with President Putin notably stating that “Whoever becomes the leader in this sphere will become the ruler of the world.” Given the wide proliferation of AI, it’s unlikely that some technology breakthrough alone will provide a sustainable advantage. Instead, success depends on combining concepts with technology.
To date, much of the focus in fielding AI and autonomous systems have focused on improving current ways of operating. Mosaic Warfare is an example of a current US Department of Defense effort to combine cutting edge technology in AI with new operational concepts to achieve a more enduring military advantage.
Mosaic Warfare offers a comprehensive model for systems warfare that encompasses requirements and acquisition processes; the creation of operational concepts, tactics, techniques, and procedures; and force presentations and force-allocation actions, in addition to combat operations.
This concept combines AI-enabled command and control with more disaggregated manned and autonomous units to enable the military to execute faster, more adaptable operations relative to an opponent.
Where these advantages stand to show the greatest promise is in the air domain as that is the domain that provides great perspective, freedom of action, and that is absolutely critical to control for the success of physical operations on the surface of the earth.
Thanks again for allowing me to present my perspectives, and I wish you all the best of success in the remainder of the conference, and in your respective endeavors.