Putting Cloud and AI Innovation into a Training Driven Combat Innovation Process
I recently attended a very good conference which addressed two key aspects of technological innovation, namely cloud computing and artificial intelligence.
The presentations and discussions were really very interesting, but what can get lost in the process of discussing technological innovation is how to drive effective combat innovation for a 0-5 military, not some hypothesized military force in 2030 or beyond.
I had a chance to put in front of the panel the question of how we both train to fight at the speed of light and to ensure that such a process informs technological development as much as new technologies provide new options for the forces.
OSD is organized to deliver “cylinder research” programs rather than being informed by the force evolving a kill web combat force.
I asked my first question after the 37 minute mark in the video which captured the seminar.
The USAF has really driven down on focusing on combat effects and Dr. Schneider underscored the importance of this approach. He argued that even the most complex advancing technology can be looked at within the context of effects-based outcomes as the technology is being shaped and evolved.
Essentially, any battlefield technological imperative, no matter how complex in R&D can be simply measured when integrated into major training range exercises.
Merging the issue of training ranges with ever advancing technology especially with regard to Cloud Computing and AI merged together in the current fight rather than suspending combat innovations until the perfect lab based maturity end game cycle is achieved is the core focus which a kill web force needs.
The process is to take the technology to the training ranges now and never stop the process of innovation into the evolving combat force. This is especially crucial given how central software upgradeable systems are becoming to the force.
Some very smart warriors will figure it all out for their peers and that training experience can be transferred rapidly to the overall combat force.
On May 30, 2019, the Hudson Institute Task Force on Federal IT Procurement, with sponsorship from Oracle and Microsoft, hosted a panel to discuss how the Department of Defense (DoD) can best implement cloud and artificial intelligence capabilities to gain the tactical edge in future warfare.
The Hudson announcement of the seminar was as follows:
A keynote address will be given by Dr. Alexander Kott, the chief scientist of the Army Research Laboratory, who will then join a discussion between Department of the Navy AI Lead Jeff Kojac and CSIS Associate Fellow Lindsey Sheppard. Hudson Senior Fellow William Schneider, Jr will moderate the conversation and audience Q&A.
As warfare becomes increasingly data-centric, cloud services and the AI systems they enable will be crucial to ensuring American warfighters have access to the most advanced information and analytics available.
With the award date of the DoD’s JEDI cloud coming as soon as July, the Pentagon will need to begin taking steps toward integrating cloud technology as a core component of its military operations.
Whichever company wins the JEDI project must provide capabilities that are not only robust, adaptable, and interoperable, but also built for the long-term future of conflict.
Dr. Alexander Kott
Chief Scientist, Army Research Laboratory
Lindsey R. Sheppard
Associate Fellow, International Security Program, Center for Strategic and International Studies (CSIS)
William Schneider, Jr.
Senior Fellow, Hudson Institute
Colonel Jeff Kojac
Department of the Navy Lead, Joint Artificial Intelligence Center, U.S. Department of Defense
And we earlier did an interview with Air Vice Marshal (Retired) Blackburn where he underscored how the dynamics of change can be shifted to ensure that a kill web evolving force can have a key say in the evolution of technology and to get that technology rapidly into the deployed combat force,.
What follows is that 2017 interview.
During my visit to Canberra, I had a chance to discuss with Air Vice-Marshal (Retired) John Blackburn how the training approach could be expanded to encompass and guide development.
“We know that we need to have an integrated force, because of the complexity of the threat environment will will face in the future. The legacy approach is to buy bespoke pieces of equipment, and then use defined data links to connect them and to get as much integration as we can AFTER we have bought the separate pieces of equipment. This is after-market integration, and can take us only so far.”
“This will not give us the level of capability that we need against the complex threat environment we will face. How do we design and build in integration? This is a real challenge, for no one has done so to date?”
Laird: And the integration you are talking about is not just within the ADF but also with core allies, notably the United States forces. And we could emphasize that integration is necessary given the need to design a force that can go up an adversary’s military choke points, disrupt them, have the ability to understand the impact and continue on the attack.
This requires an ability to put force packages up against a threat, prosecute, learn and continue to put the pressure on.
Put bluntly, this is pushing SA to the point of attack, combat learning within the operation at the critical nodes of attack and defense and rapidly reorganizing to keep up the speed and lethality of attack.
To achieve such goals, clearly requires force package integration and strategic direction across the combat force.
How best to move down this path?
Blackburn: We have to think more imaginatively when we design our force.
A key way to do this is to move from a headquarters set requirements process by platform, to driving development by demonstration.
How do you get the operators to drive the integration developmental piece?
The operational experience of the Wedgetail crews with F-22 pilots has highlighted ways the two platforms might evolve to deliver significantly greater joint effect. But we need to build from their reworking of TTPs to shape development requirements so to speak. We need to develop to an operational outcome; not stay in the world of slow motion requirements development platform by platform.
Laird: Our visit to Fallon highlighted the crucial need to link joint TTP development with training and hopefully beyond that to inform the joint integration piece.
How best to do that from your point of view?
Blackburn: Defence is procuring a Live/Virtual/Constructive (LVC) training capability.
But the approach is reported to be narrowly focused on training. We need to expand the aperture and include development and demonstration within the LVC world.
We could use LVC to have the engineers and operators who are building the next generation of systems in a series of laboratories, participate in real-world exercises.
Let’s bring the developmental systems along, and plug it into the real-world exercise, but without interfering with it.
With engagement by developers in a distributed laboratory model through LVC, we could be exploring and testing ideas for a project, during development. We would not have to wait until a capability has reached an ‘initial’ or ‘full operating’ capability level; we could learn a lot along the development by such an approach that involves the operators in the field.
The target event would be a major classified exercise. We could be testing integration in the real-world exercise and concurrently in the labs that are developing the next generation of “integrated” systems.
That, to my mind, is an integrated way of using LVC to help demonstrate, and develop the integrated force. We could accelerate development coming into the operational force and eliminating the classic requirements setting approach.
We need to set aside some aspects of the traditional acquisition approach in favor of an integrated development approach which would accelerate the realisation of integrated capabilities in the operational force.