AIToday

Speculative essay explores how AI could end power competition

LessWrong AI17h ago
Speculative essay explores how AI could end power competition

Key takeaway

A speculative essay proposes that artificial intelligence could prevent the kind of sustained power competition between nations or actors that has historically led to conflict—known as Thucydides traps. The argument hinges on two scenarios: either one actor grows so rapidly via AGI that it becomes an unassailable global hegemon, or competing actors use AI to negotiate binding agreements that make hard power unnecessary. Unlike past rulers or institutions, an AI singleton would face no mortality or external challengers and could theoretically persist indefinitely.

Summaries like this, in your inbox every morning.

Sign up free →

3 Key Points

  • What happened

    A post discusses the theoretical possibility that artificial intelligence could prevent what scholars call Thucydides traps—sustained competition between actors for control of the future—through two potential pathways: rapid growth leading one actor to become a dominant global power, or negotiated agreements enforced by AI systems.

  • Why it matters

    The piece suggests that unlike historical powers or institutions, an AI-based singleton (a single global authority) would not face death or external rivals and could theoretically maintain control indefinitely, fundamentally altering the dynamics of power competition that have defined human politics.

  • What to watch

    The author attributes the core idea to Buck Shlegeris and notes the analysis is explicitly speculative, grounded in discussion at the Forethought retreat rather than empirical evidence or forecasts.

Context & Analysis

The essay presents a thought experiment rather than a news event or product announcement. Its central claim rests on two mutually exclusive outcomes: either technological progress produces an artificial general intelligence (AGI) powerful enough that one actor outgrows competitors and forms a lasting hegemon, or uncertainty about competition outcomes creates incentives for all actors to negotiate binding agreements enforced by AI systems. In both cases, the author argues, the result is a global singleton—a single authority that, unlike human rulers or past institutions, cannot die and faces no external rivals. This permanence distinguishes an AI-based singleton from every historical hegemon in the author's view. The piece does not provide empirical support, modeling, or specific timelines; instead, it frames the possibility as plausible and worth considering. The idea originates with Buck Shlegeris, and the author credits discussion at the Forethought retreat as the prompt for the analysis.

FAQ

Is this based on new research or forecasts?
No. The author explicitly states the analysis is speculative, prompted by discussion with Buck Shlegeris and others at the Forethought retreat, not empirical study or prediction models.
What is a Thucydides trap?
The post does not define it, but refers to it as sustained competition between actors for control over the future—a concept the author suggests AI could help resolve.

Discussion

No discussion yet for this article

Stay ahead with AI news

Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.

Get Started Free

Free · takes 30 seconds · unsubscribe anytime

1 minute a day. The AI essentials.

200+ sources · Email / LINE / Slack

Get it free →