How AI Evolves with National Intelligence - Robert Cardillo

More Intelligent Tomorrow: a DataRobot Podcast

Robert Cardillo, former director of the National Geospatial Intelligence Agency joins Ben Taylor for this episode of the More Intelligent Tomorrow podcast.
Robert Cardillo has delivered dozens of intelligence assessments that informed vitally important decisions on US national security. In his 35 years serving the intelligence community, he advised President George H.W. Bush during Panama’s Noriega regime, President Obama during the search for Osama bin Laden, and to understand North Korea’s nuclear intentions.

As such, he has spent a lot of time in the Oval Office and the Pentagon contextualizing complicated, data-driven intelligence assessments to very important people. Conveying the data effectively, safely, and confidently often meant reporting on wide-ranging opinions among peers about what conclusions should be drawn.

In 2010, he was faced with a challenge of how to present critical data to President Obama on North Korea, about which there was wide-ranging dissent among his peers in the intelligence community. He needed to decide whether to present the data including his peers’ diverse assessments of what it meant, or to wait until his peers could come to a consensus about its meaning.

He decided not to wait. He included each different interpretation in his report. In the end, “President Obama told me, ‘I get really nervous when you all agree.’ So being willing to take on dissent is really healthy for the community.”

Today in his work as Distinguished Fellow at Georgetown University’s Center for Security and Emerging Technology continues to focus on data-driven information sharing and collaboration between those who collect intelligence and those who analyze it.

Listen in to hear Robert’s fascinating accounts of his 35-year service in the US Intelligence Community and its technological advances throughout his tenure.

Topics include:

• Is there an inherent flaw in the term “Artificial Intelligence”?
• What’s the most effective way to scale? How can you make a thing that worked for 10 work for 10,000?
• Can enforced authentication systems like Blockchain protect us from synthetic data, data manipulation, and deep fakes?
• If a spaceship crashed on earth, would satellite sensing capabilities we have today detect it? Or Is there too much data to review?