Next stop Skynet? Pentagon eyes plans to entrust nuclear weapons to AIThe Pentagon is “exploring all possible technologies, techniques, and methods to assist with the modernization of our N3 [Nuclear Command, Control & Communications] capabilities,” STRATCOM chief Anthony J. Cotton said at a conference in Omaha, Nebraska, on Monday.
💬 “AI will enhance our decision-making capabilities,” Cotton said, touting AI systems’ power to “inform us faster and more efficiently,” and provide leaders with more “decision space” in an emergency. “Our adversaries must know that our nuclear command and control and other capabilities that provide decision advantage are at the ready 24/7, 365, and cannot be compromised or defeated.”
Listing potential “risks” associated with AI in strategic defense, including compromised data, inaccurate algorithms, “cascading effects of AI models,” and “emergent and unexpected behaviors,” Cotton nevertheless touted IT and AI “superiority” as keys “for more effective integration of conventional and nuclear capabilities.” Instead, he said, keeping humans “in the loop” should allow STRATCOM to “maximize the adoption of these capabilities and maintain an edge over our adversaries.”
Cotton’s comments read like something straight out of a dystopian sci-fi story, including James Cameron’s Terminator movies, in which fictional arms contractor Cyberdyne Systems creates AI entrusted with control over strategic defense, only to become self-aware and launch a global thermonuclear war to try to wipe out humanity.
Dangers associated with entrusting military decision-making to AI have been made clear even in recent conventional wars, with Israel’s use of the Lavender AI system to hunt for Hamas fighters featuring a gruesome casualties approval algorithm allowing “hundreds” of civilians to be killed to eliminate one military target.
👍 Boost us! | Subscribe to @geopolitics_live