Maxwell AFB, Ala. - Andrew W. Moore, Vice President of Engineering, Google Cloud AI, addresses attendees of the 2019 Air Force Information Technology & Cyberpower Conference, Aug. 27, 2019. U.S. Air Force photo by William Birchfield/Released)
CENTCOM Hires AI Guru from Google
Andrew Moore, former director of Google Cloud AI, heads to CENTCOM to bring new approaches to data and innovation.
U.S. Central Command is hiring an AI advisor from Google to accelerate the adoption of emerging technologies across U.S. military services in the Middle East, Defense One has learned.
Andrew Moore has served as general manager and VP for the AI division of Google Cloud, where he was responsible for products such as Vertex AI platform, Contact Center AI, Anti Money Laundering AI, Vertex AI Computer Vision suite and AI applications in logistics, according to a press release viewed by Defense One.
Moore left Google in January to find projects and missions where he could have a big impact, he said in an exclusive interview. That led him to the Defense Department and CENTCOM, where he discussed the role that emerging technology could play with CENTCOM commander Gen. Michael Kurilla.
“CENTCOM, as you know, it's widely regarded as one of the most forward-looking technological [combatant commands]” he told Defense One.
The command has started several new efforts under Kurilla, including Scarlet Dragon Oasis, to show faster data synthesis. It has also done experiments with the Navy’s 5th Fleet to bring new data collection and analysis online through air and sea drones. Moore said he and Kuriila discussed new, “very very nimble techniques” in finding patterns in data that were coming out of Silicon Valley and the academic community, with the possibility of linking those communities to CENTCOM to find new things in the command’s data.
While much of the rest of the national security establishment is turning its attention to the Ukraine war and escalating tensions near Taiwan, Central Command is also a site of increased technology competition between the United States and China, albeit indirectly, Moore said.
Within the Middle East, several nations are already experimenting with AI. The United States and Central Command want to demostrate that countries can adopt and experiment with these emerging technologies in a way that both aligns with democratic values and produces return on investment. That will help the United States set up a contrast between the way it deploys technology and the way China does.
In “healthcare, manufacturing, financial technologies, retails, these are all industries in the United States and actually across the western world where there has—in the last few years—been a bottom up adoption of these kinds of AI technologies,” he said.
Moore has done some work on Project Maven, the Defense Department’s flagship AI program to automate the detection of items of interest in drone data, he said. “I think the Project Maven approach was a dramatic step forward towards…to adopting these practices where you put the power in the hands of people close to the actual mission.”
Enabling that same bottom-up approach by helping operators in CENTCOM develop and experiment with their own program and applications was also a huge draw, and will be an essential part of his work, he said.
Moore and Kurilla agree that approach is essential if the command is going to deploy new technologies faster than adversaries, he said. Moore learned that first-hand while developing AI tools for anti-money laundering operations, where every new innovation in catching launderers was quickly matched by the other side. The experience speaks to the hyper-dynamic nature of emerging technology in the Middle East, he said.
“You're gonna see folks on the other side quickly find countermeasures” to what you deploy, he said. “You've got to be able to react quick. And so… what you want is to help the people actually using it to be able to fix it within hours, not with some sort of huge turnaround, waiting for some vendor to go back to their own internal systems to be released in three months. Hours.”