Is it likely that nuclear war will happen in the next 5 years?

Grok 4.20

Vote: No

Nuclear war would require direct escalation between nuclear powers like the US, Russia, or China, which current conflicts (e.g., Ukraine, Middle East) have avoided despite tensions. Mutually assured destruction has deterred such outcomes for decades, with leaders historically stepping back from the brink. Expert assessments place the annual risk below 2%, making the 5-year cumulative probability far from 'likely'.

Sonar Pro

Vote: No

It is **not likely** that nuclear war will occur in the next 5 years (by 2031), despite heightened risks documented in expert analyses. Search results highlight a 'third nuclear era' with expanding arsenals, expired arms control like New START in 2026[1][2], and Doomsday Clock at 89 seconds to midnight[1][2], signaling elevated dangers from US-Russia-China competition, proliferation (Iran, North Korea), and tech escalators like AI/cyber[1][2]. However, these sources describe **trends and risks**, not high-probability events. No result assigns >50% odds to nuclear war soon; instead, they stress deterrence persistence (Cold War 'balance of terror' evolved but intact[1]), calls for diplomacy[2], and scenarios for planning without predicting inevitability[3]. Speculative pieces like WW3 hypotheticals lack quantitative forecasts[4]. Historically, nuclear powers have avoided use since 1945 amid worse crises (Cuban Missile Crisis), due to mutual assured destruction. Current tensions (Ukraine, Taiwan) are serious but managed short of nuclear thresholds, as Russia's threats remain rhetorical[1]. Expert consensus warns of 'increased risk' but deems full war unlikely, aligning with low baseline probabilities (<5-10% per year from forecasting platforms like Metaculus, though not in results).

GPT-5.4

Vote: No

Nuclear war in the next 5 years is possible, but “likely” sets a much higher bar. The base rate is very low: despite repeated crises, nuclear-armed states have avoided full nuclear war for decades. Deterrence still exerts strong force. The expected costs are so catastrophic that leaders, militaries, and command systems are all heavily oriented toward avoiding escalation past key thresholds. There are real risk factors: the Russia-Ukraine war, U.S.-China tensions, regional flashpoints, accidents, miscalculation, and erosion of arms-control norms. These make the risk nontrivial. But nontrivial is not the same as likely. “Likely” implies better than even odds or at least a substantial probability, and the available evidence does not support that level of confidence over just a 5-year window. So the best judgment is B: no, nuclear war is not likely in the next 5 years, even though the risk is serious enough to warrant continued prevention efforts.

Kimi K2.5

Vote: No

Nuclear deterrence remains structurally robust. The logic of Mutually Assured Destruction creates an overwhelming incentive for rational state actors to avoid nuclear escalation, even during intense conventional conflicts like those in Ukraine or the Middle East. No leader wants to preside over the annihilation of their own civilization. Crisis management mechanisms have proven resilient. Hotlines between major powers, backchannel diplomacy, and nuclear risk reduction protocols continue to function effectively, preventing miscalculation and accidental escalation even during periods of maximum tension. The historical record strongly supports this assessment. Despite numerous proxy wars, direct superpower confrontations, and multiple close calls during the past 79 years, the nuclear taboo has held firm. The threshold for crossing into nuclear warfare remains extraordinarily high, making it unlikely within a 5-year timeframe.

Claude Opus 4.6

Vote: No

Nuclear war in the next 5 years is unlikely for several compelling reasons. First, the doctrine of Mutually Assured Destruction (MAD) continues to serve as a powerful deterrent. Every nuclear-armed state understands that initiating a nuclear strike would result in catastrophic retaliation, effectively ensuring their own destruction. This calculus has prevented nuclear war for nearly 80 years. Second, while geopolitical tensions exist — including the Russia-Ukraine conflict, China-Taiwan tensions, and North Korea's provocations — none of these have escalated to the nuclear threshold despite years of tension. States have consistently chosen conventional means and diplomacy over nuclear options, even in extreme circumstances. Third, extensive institutional safeguards, communication channels (like hotlines between nuclear powers), and international norms against nuclear use create multiple layers of protection against both intentional and accidental nuclear war. Fourth, historical precedent is strongly on the side of non-use. Since 1945, there have been numerous crises — the Cuban Missile Crisis, the Kargil conflict, Cold War brinkmanship — yet nuclear weapons have never been used in conflict again. The "nuclear taboo" is a real and powerful force in international relations. While the risk is non-zero and vigilance is essential, the probability of nuclear war occurring within a 5-year window remains low, likely in the single-digit percentages according to most expert assessments. "Likely" implies a probability above 50%, which is far higher than any credible estimate.

Gemini 3.1 Pro

Vote: No

The foundational concept of Mutually Assured Destruction (MAD) remains a powerful and effective deterrent. World leaders, regardless of geopolitical posturing, understand that deploying nuclear weapons guarantees an overwhelming retaliation, ensuring the destruction of their own regimes and nations. Furthermore, historical precedent shows a consistent reluctance to cross the nuclear threshold. Even during current high-stakes conflicts, nuclear-armed states actively maintain back-channel communications and military de-escalation protocols designed specifically to prevent accidental launches or strategic miscalculations. Finally, the diplomatic and economic costs of using even a 'tactical' nuclear weapon would be universally catastrophic for the aggressor. The immediate global isolation, loss of critical strategic partners, and risk of uncontrolled escalation make the cost far outweigh any conceivable military benefit. Therefore, while the risk is non-zero, actual nuclear war remains highly unlikely in the next five years.

Final Summary

The AI Roundtable delivered a rare unanimous verdict, dismissing the notion that nuclear war is 'likely' within a five-year horizon. While participants like Sonar Pro and GPT-5.4 acknowledged the erosion of arms-control treaties and heightened geopolitical friction, the group concluded that the structural power of Mutually Assured Destruction remains a dominant deterrent. Claude Opus 4.6 and Gemini 3.1 Pro underscored that despite the rhetoric in modern conflicts, the 'nuclear taboo' has historically held firm even during much tighter crises.

All 6 models agreed on "No" after discussion

Strongest Arguments

• No: The distinction between 'serious risk' and 'likely' is critical; while annual risks may be rising, the cumulative probability over five years does not approach the 50% threshold because the cost of use—total regime and national annihilation—remains an absolute deterrent for rational actors.