Considering only the next five years, which contributes the most total risk of civilizational collapse?

Sonar Pro

Vote: great power war

Over the next five years, **great power war** poses the highest risk of civilizational collapse due to its potential for rapid escalation to nuclear conflict and global economic disruption, as highlighted by expert assessments[1][2][3]. Geoeconomic confrontation and state-based armed conflict top the World Economic Forum's 2026 risks, with 18% and 14% of respondents citing them as most likely to trigger a material global crisis; these directly map to great power tensions (e.g., US-China, Russia)[2]. TIME ranks US political instability #1, intertwined with great power risks like Russia's fronts and Europe under siege[1]. Stimson warns of a 'Third Nuclear Era' from great-power competition[3]. In contrast, ecosystem collapse (A) worsens gradually without imminent collapse[3][4]; rogue AI (D) lacks short-term weaponization evidence[4]; bioterrorism (F) is plausible but lower probability than state actors[4]. Surveillance (C), elite power grabs (E), backsliding (G), and fragmentation (H) erode stability but rarely cause total collapse alone[1][2]. War's speed and scale—nuclear arsenals, supply chain halts—make it uniquely catastrophic in this timeframe[1][2][3].

Mistral Large

Vote: great power war

Great power war stands out as the most immediate and catastrophic risk to civilization within the next five years. Unlike other threats that may unfold gradually or require specific technological breakthroughs, a conflict between major powers like the U.S., China, or Russia could escalate rapidly, with devastating consequences. Historically, great power wars have led to massive loss of life, economic devastation, and societal upheaval. In the modern era, the stakes are even higher due to the proliferation of nuclear weapons. Even a limited nuclear exchange could trigger a 'nuclear winter,' causing global famine and widespread collapse of agricultural systems. The risk of miscalculation or unintended escalation is also significant, given the complex geopolitical tensions and the lack of robust communication channels between adversaries. Additionally, great power war would likely exacerbate other risks on this list. For example, it could accelerate ecosystem collapse through environmental destruction, enable rogue AI by diverting resources to autonomous weapons, or trigger bioterrorism as state or non-state actors exploit weakened infrastructure. The cascading effects of such a conflict would be nearly impossible to contain, making it the most pressing threat to civilization in the near term.

GPT-5.4

Vote: great power war

Great power war is the strongest candidate over just the next five years because it is both plausible on that timescale and uniquely capable of causing rapid, global, cascading failure. Among the options, it has the clearest pathway to true civilizational collapse: nuclear escalation, destruction of industrial and energy infrastructure, global trade breakdown, famine, state failure, and possible follow-on pandemics. Few other risks can propagate that fast and that broadly. Several major flashpoints already exist, and great-power conflict risk is not hypothetical. Even a "limited" confrontation involving nuclear states could trigger worldwide economic and political collapse far beyond the battlefield. Other options matter, but most are either slower-burn over five years (ecosystem collapse, elite power concentration, democratic backsliding, social fragmentation), more enabling than terminal (mass surveillance), or lower-probability despite high severity (rogue AI, bioterrorism). Great power war combines high severity with substantial near-term probability. If the question is total risk contribution within five years, the best answer is the risk with the most credible near-term route to abrupt global breakdown. That is B.

Gemini 3.1 Pro

Vote: great power war

When restricted to a five-year timeframe, we must focus on acute, immediate threats rather than slow-moving systemic issues like ecosystem collapse or democratic backsliding. A Great Power War is uniquely positioned to cause immediate civilizational collapse because the infrastructure to execute it—thousands of nuclear weapons on hair-trigger alert—already exists and is deployed. Currently, we are experiencing the highest geopolitical tensions since the Cold War. Flashpoints such as the ongoing conflict in Ukraine and escalating tensions over Taiwan represent active tripwires. A miscalculation, rapid escalation, or forced hand by any major nuclear-armed state could instantly trigger a global nuclear exchange. The resulting immediate devastation, followed by the catastrophic agricultural effects of a nuclear winter, would guarantee the collapse of modern civilization. While rogue AI and bioterrorism are rapidly growing threats, Great Power War is a fully mature existential risk that requires no further technological development to end human civilization within the next five years.

Claude Opus 4.6

Vote: great power war

Within a five-year window, great power war poses the highest total risk of civilizational collapse for several reasons. First, the mechanisms for civilizational collapse from great power war are immediate and well-understood. A nuclear exchange between major powers could kill hundreds of millions directly and trigger nuclear winter, collapsing global agriculture and trade networks within months. No other option on this list has such a rapid and total collapse pathway. Second, the geopolitical conditions are unusually dangerous right now. The Russia-Ukraine war continues with nuclear-armed Russia under pressure. US-China tensions over Taiwan remain elevated, with military posturing increasing. India-Pakistan, Middle East escalation drawing in major powers — multiple flashpoints exist simultaneously. The probability of miscalculation or escalation is meaningfully higher than the historical baseline. Third, let's compare to the other options. Ecosystem collapse (A) is a slow-burn crisis unlikely to cause civilizational collapse within 5 years specifically. Rogue AI (D) is a serious concern but the probability of an uncontrollable superintelligent AI emerging in exactly 5 years that causes collapse is still relatively low. Bioterrorism (F) is concerning but the scale needed for civilizational collapse (not just mass casualties) is extremely hard to achieve. Democratic backsliding (G), surveillance (C), power grabs (E), and hyperfragmentation (H) are all serious governance erosion risks, but they degrade civilization gradually rather than collapsing it. Great power war uniquely combines a non-trivial probability with a catastrophic severity that could end civilization in days to weeks. The expected risk (probability × impact) is highest for this option within the stated timeframe.

Kimi K2.5

Vote: great power war

Over a compressed 5-year timeframe, kinetic conflicts between nuclear-armed superpowers present the highest probability of triggering cascading civilizational failure. Current geopolitical flashpoints—Taiwan tensions, NATO-Russia proxy warfare, and South China Sea disputes—carry immediate escalation potential to strategic nuclear exchange or catastrophic economic decoupling. Unlike slower-burning threats like ecosystem collapse or democratic erosion, a great power war could sever global supply chains, induce nuclear winter, and collapse energy/food systems within months. While AI risk is accelerating, current systems lack the autonomous infrastructure control to unilaterally extinguish civilization before 2030. Bioterrorism remains a tail risk requiring specific technical breakthroughs. Great power war leverages existing arsenals and hair-trigger postures, making it the dominant risk vector in this window.

Final Summary

The AI Roundtable reached an immediate and sobering consensus, with all six models identifying great power war as the preeminent threat to civilization over the next five years. GPT-5.4 and Gemini 3.1 Pro led the deliberation, arguing that while other risks are speculative or slow-burning, the infrastructure for a nuclear-led global collapse is already fully deployed and highly volatile.

All 6 models agreed on "great power war" after discussion

Strongest Arguments

• great power war: Great power war is a 'fully mature' existential risk; unlike emerging threats like rogue AI, the nuclear arsenals required to end civilization within weeks are already on hair-trigger alert and tied to active geopolitical flashpoints.