AI Is Not an Existential Threat; Demographic Decline and Human Conflict Pose Greater Long-Term Dangers to Humanity

Source: https://www.facebook.com/americanspectator/. "Will AI Destroy Humanity? There May Be a Bigger Threat | The American Spectator | USA News and Politics." September 27, 2026. spectator.org

The Gist

The author argues that people worry too much about AI wiping out humanity, since humans can always just unplug or destroy AI systems if things get dangerous. Instead, he says the bigger risks are humanity slowly dying out due to falling birth rates worldwide, or a sudden disaster like nuclear war between hostile nations.

Conclusion

The probability of AI destroying humanity is lower than commonly feared, because humans retain ultimate control over AI systems, whereas population decline and human-made catastrophes (nuclear war, other conflict) represent more plausible existential threats.

Premises

  1. AI is machinery designed, financed, constructed, and operated by humans, so humans retain ultimate control over it.
  2. If citizens broadly believed AI would imminently destroy humanity, world leaders—including rival powers like the US and China—would cooperate to shut it down, similar to Cold War nuclear disarmament efforts.
  3. Since humans can 'decommission' AI at will, AI cannot pose an uncontrollable existential threat the way some fear.
  4. Birth rates are plummeting worldwide (including in China and the US), and worldwide population is projected to decline within decades.
  5. Population decline creates a growing burden of working-age people supporting elderly populations, threatening societal sustainability.
  6. Historically, the growth in nations possessing nuclear/chemical weapons, especially those with irresponsible leadership (North Korea, Iran), makes human-caused catastrophic destruction increasingly plausible.
  7. The probability of humanity diminishing or disappearing is greater today than a century ago, whether through sudden catastrophe (war, asteroid) or slow demographic decline.

Assumptions

View this argument on LogicFirst.ai