Five Civilizational Risks Converging on the Same Decade
AI risk. Nuclear risk. Pandemic risk. Climate risk. Biodiversity collapse. They're not independent — they interact, amplify each other, and are all accelerating toward the same window: 2030-2050.
5Major civilizational risk categories tracked
2030-50Critical convergence window
~30%Estimated total catastrophic risk this century (Ord)
9Planetary boundaries — 6 now breached
Choose your depth. The data doesn't change — just the explanation.
Scientists who study things that could go very wrong for all of humanity — like AI going out of control, nuclear war, pandemics, climate change, and animals going extinct — notice something scary: all these big risks seem to be getting worse at the same time. And they don't just exist separately — a climate disaster might cause more wars, which might increase the chance of nuclear use. This is called "compound risk" — risks that make each other worse.
Existential and civilizational risks are threats that could permanently curtail humanity's potential. The main candidates: unaligned AI (10% this century, Toby Ord), engineered pandemic (10%), nuclear war (1%), climate change (1%, though impacts interact with others), and biodiversity collapse (part of climate risk). Critically, these risks interact: climate stress destabilizes governments (nuclear risk), compromises institutional resilience to pandemics, and creates the resource pressures that AI-enabled surveillance states may exploit. The 2030s are when multiple risk curves peak simultaneously.
Ord "The Precipice" (2020): 1/6 (~17%) total existential risk this century. Distribution: AI 10%, engineered pandemic 10%, nuclear war 1%, climate <1%, other <1% — but interaction effects not modeled. Planetary boundaries (Rockström et al., 2023 Science Advances): 6 of 9 breached (biosphere integrity, climate change, land system change, freshwater, biogeochemical flows [P,N], novel entities including plastics). Doomsday Clock: 90 seconds to midnight (2023) — closest ever. Compound risk literature: Kemp et al. (2022) "Climate Endgame" argues 4°C+ paths interact with conflict, disease, food security. NTI biosecurity index 2023: pandemic readiness declining. Compound AI-nuclear risk: AI-enabled early warning systems could reduce or increase accidental nuclear launch probability depending on implementation. Feedback loops: economic disruption from AI → political instability → nuclear doctrine loosening.
Ord "The Precipice": theprecipice.com. Planetary boundaries Rockström 2023: science.org/doi/10.1126/sciadv.adh2458. Doomsday Clock: thebulletin.org/doomsday-clock. Kemp et al. "Climate Endgame" PNAS 2022: pnas.org/doi/10.1073/pnas.2108146119. Global Catastrophic Risk Institute: gcrinstitute.org. NTI Biosecurity Index: ntiindex.org. Future of Life Institute: futureoflife.org. Nuclear Risk Reduction Center: nrrc.state.gov. IPBES biodiversity assessment: ipbes.net.
The Composite Risk Dashboard
Individual Risk Probability (% this century) — Expert Estimates
Estimated probability of each risk causing civilizational catastrophe
Risk Trajectory — Worsening vs. Improving (2000-2024)
Composite risk index by category (0=low, 10=extreme)
Risk Interaction Matrix — How Each Risk Amplifies Others
Amplification score: how much each risk increases others (0-3 scale)
🔗 Why the Interaction Effects Are What Matters
The naive view: each risk has its own probability, they're independent, total = sum. The realistic view: they're deeply coupled. Climate change stress on food/water systems increases migration and conflict, which strains the international institutions needed to control AI and nuclear weapons. A pandemic that kills key technical staff could accelerate unsafe AI development as companies race to recover productivity. An AI arms race creates pressure to cut corners on safety in exactly the period when the technology is most dangerous. Toby Ord's estimate of 1-in-6 overall existential risk assumes some independence. If the interaction effects are large, the real number could be significantly higher. Nobody has a good model for this — which is itself a civilizational risk.