By: Margaret Jane Piatos
American companies have spent $335 billion and counting on artificial intelligence (AI) between the years 2013 and 2023, and today, 79% of Americans interact with AI several times a day. What began as a niche technological field has quickly become embedded in finance, healthcare, education, defense, and even personal decision-making. As governments and corporations race to dominate this emerging technology, the pace of development far outstrips that of regulation.
AI is as dangerous and as strategically transformative as nuclear weapons were in the 20th century, and it must be governed with the same urgency and international cooperation before competition turns into catastrophe.
In the 20th century, nuclear weapons transformed international politics almost overnight. They redefined how we fought wars, and consequently, who held hegemony. But nuclear technology also forced the world to confront a terrifying reality: innovation without restraint could mean total destruction.
The atomic bomb dropped on Hiroshima in 1945 killed an estimated 100,000 people. Modern tactical nuclear weapons are far more powerful, and at the height of the Cold War, the global arsenal grew large enough to destroy human civilization many times over. It took regulations such as the 1968 Nuclear Non-Proliferation Treaty, the International Atomic Energy Agency, and Cold War arms-reduction agreements like SALT and START, to stabilize the system.
Today, artificial intelligence is advancing even faster — and with far fewer guardrails.
Unlike nuclear weapons, AI does not require rare uranium or massive reactors. However, it requires data, computing power, and engineering talent. That makes it easier to develop and more difficult to contain. Yet its strategic impact may be just as profound. AI can enhance cyberwarfare, power autonomous weapons systems, manipulate financial markets, disrupt infrastructure, and more. The state that leads in AI will not just have faster software, but it will have a structural advantage in economic productivity, military capability, and information control.
And countries know it.
The United States and China are locked in a race for AI dominance. Washington frames AI leadership as critical to national security and economic competitiveness. Although the United States still maintains a far larger nuclear arsenal with 5,277 warheads compared to China’s 600 warheads, today's balance of power may not be determined by stockpile size alone.
Beijing has openly declared its ambition to become the world leader in AI, embedding AI development into its long-term industrial planning and military modernization strategy. In some areas, particularly the integration of AI into surveillance systems and military deployments, China may be moving faster and with greater coordination than Western approaches.
The European Union, meanwhile, is racing to shape regulatory standards that could define global norms. Through its landmark AI Act, the EU has moved to classify AI systems by risk level, impose strict transparency requirements on high-risk applications, and ban certain uses such as social scoring and some forms of biometric surveillance. By prioritizing safety and human oversight, the EU is attempting to export its regulatory model beyond Europe’s borders. If successful, the EU could set the rules of the road for AI governance worldwide, not by dominating innovation, but by defining the standards companies must meet to access one of the world’s largest markets.
This competition resembles the early years of the arms race, but with one critical difference: AI development is primarily driven by private companies operating across borders. Governments are competing not only with rivals abroad but also, motivated by President Trump’s AI challenge, racing to outpace domestic innovation.
That dynamic creates a dangerous incentive structure. When speed becomes the priority, safety becomes secondary. The result is an emerging AI arms race without clear rules or shared red lines.
The risks are not hypothetical. Autonomous weapons systems could make battlefield decisions faster than human oversight allows. AI-generated misinformation could undermine public trust in democratic institutions. Advanced cyber capabilities could cripple infrastructure without a single missile being launched. And as AI systems grow more complex, even their creators may struggle to predict their behavior.
Nuclear weapons taught the world a painful lesson: revolutionary technologies cannot be managed through optimism alone. Governance must evolve alongside capability.
AI requires a similar mindset. It means establishing guardrails before chaos forces them upon us. Nations should cooperate to limit fully autonomous lethal weapons, prevent AI from controlling nuclear launch systems, and establish monitoring mechanisms for the most advanced AI models. Export controls on advanced semiconductor chips are already a step toward recognizing AI’s strategic importance, but they are only the beginning.
Some argue that AI is too embedded in civilian life and too rapidly evolving to be regulated like nuclear weapons. They are right that AI is different. But that makes regulation more urgent. And we must treat it as such, before competition becomes a catastrophe.
