Short answer

• The biggest plausible existential threats (risks that would permanently destroy or permanently and drastically curtail humanity’s long-term potential) are: misaligned advanced artificial intelligence (AI), engineered pandemics and other catastrophic biological risks, nuclear war (especially escalation to global nuclear winter), severe runaway climate/ecosystem collapse that destroys civilization’s capacity to recover, advanced destructive nanotechnology, and — at much lower probability but still possible — asteroid/major volcanic events. Combinations and cascading systemic failures increase overall danger. Unknown unknowns (“surprises”) are also important.

• Humanity can substantially reduce these risks by: reducing the chance of catastrophic events (safer design, secure arsenals, better lab and tech governance), improving detection and early response systems, investing in robust recovery and resilience, and creating international institutions and norms that coordinate responses. For AI specifically, focused alignment research, verification, governance, and deployment controls are crucial.

• How likely are we to overcome them? Uncertain. Expert judgments range widely. A careful recent estimate of existential risk this century is on the order of a few percent to a few tens of percent (estimates like ~1 in 6 appear in the literature), but these numbers are highly uncertain and depend on technological pathways and policy choices. With serious global investment in risk reduction, we can plausibly reduce the chance substantially.

Longer answer — main threats, how they can be overcome, and rough odds (large uncertainty)

1. Misaligned advanced AI

• What it is: Future AI systems that are extremely capable but whose objectives or behavior are not aligned with human values and control, potentially producing catastrophic outcomes (loss of control, rapid destructive optimization).
• How to overcome it:
  • Invest heavily in AI alignment research (theory and tools for interpretability, verification, robustness).
  • Slow or stage deployment of capabilities until alignment can be reliably demonstrated; use governance to require safety audits, red-teaming, and oversight.
  • Create multi-stakeholder international norms, capability monitoring, and coordinated responses for rapid capabilities.
  • Build technical and institutional “tripwires” and containment tools (sandboxing, compute and model provenance tracking).
• Rough chances this century: highly uncertain. Expert estimates range from very small to substantial (order-of-magnitude ranges like <1% up to tens of percent for catastrophic/ existential outcomes). Many researchers treat this as one of the highest-risk areas precisely because of the potential rapidity and scale of impact.

2. Engineered pandemics and catastrophic biological risks

• What it is: Deliberate or accidental creation and release of pathogens with high transmissibility and lethality or novel bioagents that bypass defenses.
• How to overcome it:
  • Strengthen global surveillance, rapid diagnostics, and sequencing networks to detect outbreaks early.
  • Invest in rapid-response vaccine and therapeutics platforms, stockpiles, and distribution infrastructure.
  • Improve laboratory biosafety and biosecurity standards, regulation of synthesis and distribution of genetic material, and norms against misuse.
  • International cooperation on information sharing, response coordination, and strong penalties for misuse.
  • Research into broad-spectrum countermeasures and platform technologies (e.g., antiviral platforms) and decentralized public-health capacity.
• Rough chances this century: non-negligible. Natural pandemics are unlikely to cause total extinction but engineered pathogens could be much more dangerous. Expert ranges for existential-level biological risk in the next century are roughly 0.1% to a few percent (higher in some expert elicitation scenarios) — very uncertain and strongly dependent on governance.

3. Nuclear war (and nuclear winter)

• What it is: Large-scale nuclear exchanges could kill millions directly and, by launching soot into the atmosphere, cause a “nuclear winter” reducing food production worldwide and risking societal collapse.
• How to overcome it:
  • Arms control agreements, verified reductions in arsenals, de-escalation policies, and crisis-management mechanisms.
  • Reducing hair-trigger alert postures and insecure command-and-control practices.
  • Strengthening diplomatic channels and conflict prevention.
  • Food system resilience planning for potential shocks.
• Rough chances this century: probability of global nuclear extinction is low but non-negligible for civilizational collapse. Estimates vary, but many researchers put the risk of catastrophic global consequences from nuclear war at a few tenths of a percent to a few percent this century depending on geopolitical tensions.

4. Runaway climate change and ecological collapse

• What it is: Severe warming and associated tipping points that could undermine agriculture, infrastructure, and social order. Climate by itself is unlikely to cause complete extinction in the near term, but could trigger cascading collapses, mass suffering, and loss of civilization’s ability to recover and thrive.
• How to overcome it:
  • Rapid, deep reductions in greenhouse-gas emissions (energy transition, efficiency).
  • Carbon dioxide removal, adaptation measures for infrastructure and food systems.
  • International cooperation on climate finance, technology transfer, and loss-and-damage mechanisms.
  • Preserve biodiversity, protect ecosystem services, and reduce pressures (land use change, pollution).
• Rough chances this century: high chance of serious damage and large suffering; lower chance of outright extinction. Probabilities for existential outcomes are debated; many estimate climate-alone existential risk this century is small (<1%) but the chance of collapse/regional catastrophic outcomes is much higher.

5. Advanced nanotechnology / molecular manufacturing

• What it is: Hypothetical future capabilities to manufacture harmful materials or devices at scale (weaponized self-replicating systems, extremely cheap means to create destructive hardware) — “grey goo” is a popular but contested scenario.
• How to overcome it:
  • Responsible research and development, regulation, access controls, monitoring of critical manufacturing technologies.
  • Robust safety-by-design principles and international norms.
  • Detection systems and rapid-response capabilities to contain misuse.
• Rough chances this century: very uncertain and dependent on whether such technologies materialize in the dangerous forms predicted. Many experts treat this as a moderate-to-low probability but high-impact risk.

6. Asteroid impacts and supervolcanic eruptions

• What it is: Large asteroid/comet impacts or massive volcanic eruptions could produce global catastrophes.
• How to overcome it:
  • Telescopic surveys and early detection of near-Earth objects.
  • Invest in deflection and mitigation capabilities, evacuation and food storage plans.
  • Improve monitoring of volcanic activity and contingency planning.
• Rough chances this century: very low for truly extinction-level events (<0.01%); we can reduce them further with detection and preparedness.

7. Systemic and societal risks / social collapse

• What it is: Political fragmentation, loss of institutions, prolonged global supply-chain failure, AI-enabled misinformation or cyberattacks leading to civilization breakdown.
• How to overcome it:
  • Build societal resilience: diversified supply chains, local food production, robust cyberdefenses, strong institutions, civic education.
  • Foster social trust, reduce inequality and polarization, improve global governance and accountability.
• Rough chances this century: non-negligible for localized or regional collapses; global-scale permanent collapse is harder to estimate but is a meaningful possibility in scenarios where multiple risks interact.

Cross-cutting solutions that reduce many risks simultaneously

• Better global governance and coordination: treaties, inspection regimes, verification, rapid crisis diplomacy.
• Robust early-warning systems and monitoring across domains (biosurveillance, AI capability monitoring, missile detection, climate tipping indicators).
• Invest in resilience and redundancy: distributed energy, food storage, decentralization of critical infrastructure, knowledge preservation.
• Scientific investment in safety: AI alignment labs, biosecurity research, public-health infrastructure, and defensive technologies.
• Norms, incentives, and accountability for peaceful use of powerful technologies; better education and public engagement.
• Insurance, contingency planning, and exercises that stress-test global systems.

What individuals and institutions can do (practical steps)

• Support and advocate for policies that reduce structural risks: arms control, climate mitigation, biosecurity, AI safety funding.
• For scientists and companies: adopt best practices for safety, transparency, and external review.
• Invest in community and local resilience: disaster preparedness, civic engagement, building social capital.
• Donate or work for high-impact organizations focused on risk reduction (biosecurity, AI safety, climate mitigation, arms control).

Overall chances and uncertainty

• All probability numbers above are highly uncertain; expert elicitation studies and books like Toby Ord’s The Precipice suggest non-trivial existential risk this century (his rough central estimate ~1 in 6), but other researchers give lower or higher numbers. Reasonable, policy-relevant interpretations put the total existential-risk-of-the-century somewhere between a fraction of a percent and a few tens of percent, depending on technological trajectories and choices humanity makes in governance and investment.
• Crucially: these are not fixed. Governance, safety research, and international cooperation can substantially lower risk; conversely, reckless deployment and poor regulation can raise it. Many of the most severe threats are created by human technology and so are, in principle, controllable.

Bottom line

• The existential threats are real, varied, and often interlinked. The most important practical takeaway is that relatively modest investments in governance, safety research, surveillance, and resilience now can dramatically lower the odds of catastrophe later. Humanity’s chance of surviving and flourishing is not guaranteed, but with coordinated effort the risks can be kept low enough that a thriving future remains overwhelmingly plausible.