When the Department of Defense rolled out the Artificial Intelligence Strategy in 2018, the Joint Artificial Intelligence Center (JAIC) was at its core and established the same year. The Naval Postgraduate School is hosting the JAIC’s director, Lieutenant General Jack Shanahan, U.S. Air Force (Retired); and chief operating officer, Mr. Nand Mulchandani, on Tuesday, 13 October to address the topic “Harnessing AI for National Security.” Ahead of this Secretary of the Navy Guest Lecture series event, Lieutenant Commander Karen Kutkiewicz, U.S. Coast Guard and U.S. Naval Institute Executive Fellow, interviewed Mr. Mulchandani.
Proceedings: What is the role of Joint Artificial Intelligence Center? How does it interact with the services?
Mulchandani: The JAIC (pronounced “Jake”) is a center for excellence for AI. The DoD established the JAIC in 2018 and Lieutenant General Shanahan was the plankowner. I joined the team about six months later. Our primary mission initiatives benefit America’s national security, and our servicemembers. My background is as an entrepreneur in the Silicon Valley. The JAIC is a fusion between the U.S. Military and Silicon Valley. This level of teaming together is unique, with tech and military.
Our goal is to deliver AI-enabled capabilities at scale that transforms the DoD for the digital era. We use the best of what is there with the best of new ideas and thinking and merge them together. It is two-sided. On one hand, we know that new technology is rapidly evolving. Just like in the tech industry, there are pathfinders that try everything out, that is our job for the DoD. What new technologies would most benefit us? On the other hand, we get a demand signal from our customer, whether it is a need for business process transformation, innovative health solutions, predictive maintenance, or AI-enabled warfighting capabilities. At the JAIC, our job is to survey the demand side, bring the tech side, and create a solution.
There are three million members in DoD. At the JAIC, we employ 180 people. We cannot drive change ourselves. We need everyone in the DoD to bring their needs to light. We leverage our tech background and thinking. The JAIC’s job is to not make something once. Highly successful business models are leveraged, repeatable, and scalable. That is how we bring AI capability to the entire organization. We are building platforms open to all users of the DoD, and operating at an industrial scale. We research, test, prototype, mechanize, evaluate, and package AI, all for the precise use of our Armed Forces.
Our relationships with the services and combatant commands continue to grow and thrive, and we think of them as our customer base. We have a mission team, and their job is to engage and interact with the combatant commands and the services to understand the command demand signals side of the equation. As we are building out the relationships, we are learning from them what their needs and requirements are to execute their mission. We have service representatives and subgroups for AI policy, testing and evaluation, ethics, education and training, and AI governance. We find common elements to the problems they are having and develop generic solutions that are customizable for their particular use case. The heart of our customer engagement mission is to learn the service’s needs and requirements, take our technological expertise, to create efficiencies for American competitive advantage.
Proceedings: What advance/s are you most excited to see come to fruition?
Mulchandani: My aspirational vision is the ability of us as a department to develop crucial software that will be the key determinate to victory in a future fight. This software will be our literal ability to access and compute data to build more advanced AI products. Software for predictive maintenance will predict failure or success in the field. In the next couple of months, we will be seeing an early version of the breakthrough ‘Joint Common Foundation’, to enable servicemembers at scale to use the technology that we make. The Joint Common Foundation is a cloud-enabled AI platform that will build momentum across the entire DoD enterprise. It will accelerate AI development, testing, and evaluation of new capabilities. General Jack Shanahan has a great analogy for this, the department of electricity. There is no department of electricity anymore but there are electricians. When we do our jobs right, the JAIC will leave behind only power stations and electricians. We will have AI integrated into the fabric of what we do as an organization. This improves the DoD, and that is the win.
Proceedings: Does China or Russia have an advantage in AI? If yes, what are their advantages and what can we do about it?
Mulchandani: First, AI is not a single technology. It is a collection of algorithms and modeling. AI used to be called statistics- a collection of technology, string, and numerical data to make predictions. Saying China or Russia is very broad. One country is too wide to be characterized. Tech matures and grows where we focus, where we invest our manpower and resources. There are certain fields where each country is more heavily invested. Look at the great fire wall in China, filtering on key words in searches, facial recognition for tracking of citizens. China does have the lead in certain areas that they care most about, population control, law enforcement, etc. We, as the United States, have not necessarily focused on these areas.
As Americans, we are the ones innovating, the universities conducting the extensive research, publishing, and implementing technology to scale. There is no question that United States still has the advantage. The challenge is that the Chinese work is in house and the PLA gets the first fruits. At the JAIC, we work with industry, take their first fruits, and integrate that into our system. The real challenge is that sometimes the military can be rigid. But in our short two-year existence, we are paving the way toward integration and improving the day-to-day life of each servicemember.
Proceedings: What are some ways AI will significantly change our day-to-day lives or day-to-day operations for the military?
Mulchandani: We are building a leveraged business model where 180 people hope to change a three-million-person organization. Our job is to do something so that every marginal piece of inputs or work that can affect 10,000 or 20,000, or 100,000 people. That is the leverage business model and that becomes the heart what the JAIC is aspiring which is by being a Center of Excellence by building a platform that can open to all users in the DoD. We customize the platform to them, and it grows more precise with more data and input. When we are taking about lethality and warfare, we need precision. And not just precision one time, but repeatable precision. AI is becoming woven into the fabric of our personal and professional lives. The JAIC is delivering AI-enabled solutions to our everyday problems that affect our troops around the globe.
Proceedings: Last month, the Center hosted the first-ever International AI Dialogue for Defense, what were the biggest takeaways?
Mulchandani: For me, the biggest takeaway was that we are closer to our allies than we are apart with respect to systems and ethics. We may be different culturally, but the more we engage with our partners and allies, the more we learn that we agree on most principles and ideas. The resulting dialogue strengthened our alliances. We, the U.S. and the JAIC, are the ones who are leading and sparking this discussion as the host nation. Thirteen countries came together for the AI Dialogue for Defense. Ethics and AI go hand in hand. We agreed upon and used an ethical framework to facilitate open dialogue. That was marked success.
One of the biggest challenges to solve in AI is unintended consequences and a primary topic of the Dialogue. Basically, if I can’t explain it, I can’t test it, I can’t trust it, then it would be unpredictable, and have unintended consequences. But if I can explain, test, trust, predict, then it would produce the intended consequences that we expect. Most AI in the world has been productionalized. That is, where the consequences of being wrong is very low, like Siri, Netflix, etc. You can try to change your voice or add more data points to try to fix it, but if it is wrong you watch Tiger King instead of Stranger Things. To date, AI has low consequences as we have deployed AI based cognitive assistance systems without the serious consequences. Now we are dealing with weapons systems. That is why we here at the DoD are spending so much time to ensure excellence that is aligned with America’s laws and values. The problems that we are dealing with have life and death consequences. We are even working on a Search and Rescue product for the Coast Guard. The stakes are high for us to complete our mission. What improvements we do today to accelerate and deliver AI to the warfighter is a game-changer. As I said before, this will be the difference between winning and losing whatever comes next.