The Von Neumann Gambit: AI, Espionage, and the End of Humanity?
Can we survive technology?
John von Neumann—a guy so brilliant, he made even the cleverest of us look like we were still playing with crayons. If James Bond ever needed a villain who could outthink him with a slide rule, Von Neumann would be the guy. Architect of game theory, progenitor of the modern computer, and, oh yes, key player in the Manhattan Project—the man was a walking, talking, chess-playing embodiment of what happens when sheer intellect meets a worrying lack of sleep.
Let’s picture it for a moment: Von Neumann today, feet up on an old desk in the dimly lit chaos of his Princeton office, sipping whiskey and pondering over the existential threat posed by artificial intelligence. He’d likely glance over the latest AI headlines, snort derisively, and mutter something like, “AI should be secure by design, or we’re all going to need a lot more whiskey.”
And that, dear reader, is where we should start.
Colourized photograph of von Neumann from 1954 (Photo: Courtesy of the IAS archive)
AI: The Toddler with a Flamethrower
Yuval Noah Harari confirmed in Nexus that artificial intelligence isn’t your typical gadgetry. You can’t bash it with a spanner when it misbehaves. No, it’s already out there—creeping through your phone, whispering through your fridge, and probably eavesdropping on this very remote part of the web.
Von Neumann, the mastermind behind self-replicating automata, would have seen the parallels immediately. AI doesn’t just sit idly in a box; it evolves, learns, and—if we’re careless—figures out how to drive the bus better than we can. But what if the bus is hurtling toward a cliff? Or worse, what if it thinks the cliff is precisely where we should be going?
The parallels to nuclear technology are glaring. Von Neumann’s world was one where humanity flirted with annihilation through a mushroom cloud. His work balanced on the edge of destruction and utility—a tightrope we now find ourselves walking with AI. The difference? This time, we don’t have a Manhattan Project with a singular focus.
Instead, we’ve got a bunch of toddlers with flamethrowers and no fire marshal in sight.
The Need for a Secure Foundation
Von Neumann would undoubtedly advocate for a Plan A, B, and C. But today’s tech giants? They seem content rolling out AI systems with the urgency of a coffee chain launching pumpkin spice lattes—security often an afterthought.
AI must be secure by design, with built-in safeguards from the moment a line of code is written. Think of it like designing a plane that can’t crash, rather than handing out parachutes after takeoff. Automatic updates, multi-factor authentication, and zero-trust architectures are a start. But this isn’t just about locking the doors—it’s about making sure the house doesn’t spontaneously combust.
Moving Target Defense: A Gambit for Survival
Here’s where the Von Neumann gambit might just play out. Moving Target Defense (MTD) offers a strategy that feels almost Neumann-esque. It’s about shifting the playing field, constantly changing the attack surface so that threats—human or otherwise—can’t lock on. Imagine if the AI you’re worried about hacking into your system finds itself chasing shadows instead of real targets.
It’s not foolproof, of course, but what is? The point is to keep the adversary guessing.
Von Neumann understood that in any game—be it chess, nuclear brinkmanship, or AI development—the element of unpredictability is often the strongest weapon.
And if unpredictability can buy humanity enough time to evolve smarter defenses, perhaps that’s a gambit worth playing.
The Human Endgame
At the heart of this lies a sobering truth:
AI isn’t going away, and neither are its risks.
As we delegate more decisions to machines, from credit scores to cancer diagnoses, the stakes grow higher. Von Neumann warned us about the dangers of self-replicating systems, and today we’re on the verge of seeing them play out in real time.
The question isn’t whether AI will change the world—it already has. The real question is whether we’re savvy enough to build it securely, predict its moves, and—if necessary—outplay it.
And if not? Well, I hear dolphins are excellent conversationalists.
Have a great 2025, folks.
For those interested in exploring the challenges and implications of technological advancement, John von Neumann's 1955 lecture, "Can We Survive Technology?" offers a profound perspective. In this lecture, von Neumann delves into the rapid pace of technological progress and its potential consequences on society, emphasizing the need for intelligent judgment in navigating these advancements.