Surveilling the Future of American Intelligence
Amy Zegart joins Michael McFaul on World Class podcast to talk about Spies, Lies and Algorithms, her new book exploring how the U.S. intelligence community needs to adapt to face a new era of intelligence challenges.
James Bond; Jason Bourne; Jack Bauer: cinema spies like these are the suave and daring face of spycraft and intelligence for most people.
That’s a problem, says Amy Zegart, a senior fellow at the Center for International Security and Cooperation (CISAC). In her new book, Spies Lies and Algorithms, Zegart draws on her expertise in U.S. intelligence and national security to debunk some of the pop cultural tropes around spycraft, many of which have a surprisingly pervasive influence on how both the public and policymakers understand the intelligence community.
She joins FSI Director Michael McFaul on World Class podcast to talk about the book and what she’s learned from her research about the challenges the U.S. intelligence community will need to meet in order to stay competitive in a rapidly-evolving, increasingly digital world.
Listen to the full episode now. A transcript and highlights from their conversation are available below.
Click the link for a transcript of “Spies, Lies and Algorithms with Amy Zegart."
The Origins of Spies, Lies and Algorithms
When I was a professor at UCLA, I was teaching an intelligence class. On a lark, I did a survey of my students and I asked them about their spy-themed entertainment viewing habits, as well as their attitudes towards certain intelligence topics like interrogation techniques.
What I found was there was a statistically significant correlation between their spy-themed viewing habits; people who watch the show 24 all the time were far more pro-waterboarding, among other things, than students who didn't watch spy-themed entertainment.
This got me really thinking, what do people know about espionage? Where do they get this information? And what I found is that most Americans don't know anything about the intelligence community, and when you're talking about spy agencies in a democracy, that's really problematic.
The original book was going to be more of a textbook for the class I was teaching, but as I write it, so many things started to change in the intelligence community, both politically and technologically. So with that in mind, I tried to make it forward looking to where intelligence needs to go, not just backward to where it’s been.
Who are the Spies?
The biggest surprise for me came from the research I did on open-source intelligence and nuclear threats. If ever there was an area you would think spy agencies would have cornered the market on intelligence, it would be nuclear threats.
But what I found is that there's a whole ecosystem of non-governmental people tracking nuclear threats around the world and actually uncovering really important things. That includes some open-source nuclear sleuths here at Stanford among my colleagues at CISAC.
The more I dug into this, the more I realized that open-source intelligence and publicly available data is the ballgame in the future of intelligence. Secrets still matter; but they matter a whole lot less than they did even ten years ago.
There is so much insight that we can glean from open-source information, but the intelligence community has to figure out better ways to connect with organizations and people who are in this ecosystem.
What are the Lies?
Intelligence is about deception. We don't want our enemies to understand all of our military capabilities, for example, or what our intentions are.
But I think the technology and data revolutions of the last few decades has also changed the nature of deception. It’s gone from elites deceiving elites about where their troops are, and whether they're going to attack, to mass audience deception and this disinformation-information warfare of domestic audiences like we're experiencing here in the United States.
Understanding deception, not in a pejorative sense but as an analytic frame, is critically important to being able to gather and analyze intelligence correctly.
The Power of Algorithms in Intelligence
We're drowning in data. The amount of data on Earth is estimated to double every two years. Think about that for a minute. It's just an astounding amount of data. And it's too much for any human to deal with.
So, if intelligence is in the business of collecting or finding needles in haystacks, the haystacks are growing exponentially. The intelligence community has to use artificial intelligence and other tools to augment human analysts. AI frees up analysts to then ask questions about things like intent, which humans can figure out much better than machines.
Imagine an algorithmic red team: you have humans that are developing assessments of what is Putin going to do in Ukraine, and you've got a red team that's just algorithms aggregating data, so that you have a sort of competitive analysis between humans and machines that can make the humans better. Those are the kinds of things that intelligence agencies need to be doing much more with AI.
