Spies, Lies, and Algorithms: The History and Future of American Intelligence by Amy Zegart. Amy holds a Ph.D. in Political Science from Stanford University. Dr. Zegart is an associate professor at UCLA’s School of Public Affairs.
Amy previously served on the Clinton administration’s National Security Council staff in 1993 and as a foreign policy advisor to the Bush-Cheney 2000 presidential campaign.
She has testified before the Senate Select Committee on Intelligence and has provided training to the Marine Corps, the Office of the Director of National Intelligence, and Lawrence Livermore National Laboratory.
So, it is no surprise US intelligence does not publicly address their embrace of AI for obvious reasons. Amy is documenting the use of technology including AI in the world of espionage. US intelligence has the challenge of confronting the James Bond 007 effect when confronting both public opinion and the growing role misinformation.
Amy is providing a historical view of US intelligence and their embrace of technology. She is also offering a future view of American espionage in a world of advanced AI. This is a very interesting read to discover an overview to US intelligence and the history of fatal biases and misunderstood analytics. Yet, Amy is outlining how today’s technology empowers both old 3rd world and new enemies. Technology has also empowered citizens to use web services to track nuclear threats. This was unheard of during the Cold War.
Upon a closer examination at intelligence analysis Amy reveals why simple facts are overlooked:
The answer lies in two sets of factors: natural variations in the predictability of human events and the limitations of human cognition. The world out there is inherently hard to understand and human brains are not wired to understand it well.
pg. 110
These examples trace back to George Washington’s espionage role during the Revolutionary War to the Stuxnet cyber attack.
The seven deadly sins of bias
In fact, Amy is outlining the critical role of bias within US intelligence community failures. Chapter 5: Why Analysis Is So Hard – The Seven Deadly Biases is perhaps the most important element of this book. Amy documents how General Douglas MacArthur’s confirmation bias led to the 1950 collapse of UN efforts. Repeatedly ignoring US intelligence that China would enter the war, MacArthur relayed to President Truman that China could only send 50,000 troops into the war. However on November 25th over 300,000 Chinese troops ambushed UN forces along the Yalu river. China’s entrance led to the loss of North Korea and Seoul falling in January 1951.
However by April the Joint Chiefs of Staff drafted orders for MacArthur authorizing nuclear attacks on Manchuria and the Shandong Peninsula if China launched airstrikes. President Truman arranged for the transfer of nine Mark 4 nuclear bombs to military control from the United States Atomic Energy Commission. However, the Joint Chiefs were not convinced MacArthur might prematurely carry out orders to engage nuclear weapons. President Truman would relieved MacArthur shortly thereafter.
Bias and the errors of misinformation
Amy’s analysis of how bias limited US intelligence efforts is further outlined via: Optimism Bias, Availability Bias, Correspondence Bias, Mirror Imaging, Framing Bias, and Groupthink Bias. this is a very revealing set of case studies that would benefit any organization today.
Amy references two books that I have read surprisingly almost ten years ago: Philip Tetlock’s Superforecasting and to no surprise Daniel Kahneman’s Thinking, Fast and Slow:
We certainly cannot ignore open source intelligence made available to the world via the internet. Perhaps the tweet “Helicopter hovering above Abbottabad at 1 a.m.” is proof regarding our world’s digital transformation. No doubt, the role of GoogleEarth today versus imaging via 1950’s U2 spy planes is also confirmation of our transformation. Amy draws in the closing chapter the new and still emerging digital battlegrounds in cyberspace.
Finally, Amy is able to precisely address the buzz ‘data is the new oil’ and correct a number of false assumptions:
Many have called data “the new oil” —a precious resource that fuels geopolitical competition. In many ways, it’s a poor analogy. The differences between oil and data are stark, and they offer insight into just how much the world is changing. Oil is valuable because it is scarce; data is valuable because it isn’t. Data is essentially infinite and everyone can get it—creating network effects. Oil is captive to geography, making some countries more powerful than others. Data is unbound by geography, making even powerful countries vulnerable to attack (more on that below). Adversaries cannot turn oil into water or make it look like something it isn’t. But they can with data, corrupting it or generating so much uncertainty that nobody trusts it. Data is simultaneously mighty and weak.
pg. 482
In conclusion, Amy has presented a deeply researched and intriguing overview of US intelligence embracing technology. A certainly revealing book and the very powerful role of bias impacting espionage.