No analogy for AI will hold—regulation and security need entirely new frameworks

jachiam0 · x · 2026-09-24

A researcher argues that none of the existing analogies for AI—normal technology, industrial revolution, banking regulation, nuclear arms treaties, cybersecurity, or an alien species—will hold up indefinitely. Each is seductively explanatory for some facet of AI development or impact but breaks down in another important regime.

The implication: we will eventually be forced to admit no single analogy works, and must build a genuinely new kind of conceptual, regulatory, societal, and international security framework to deal with AI, since it is too many things at once for any one analogy to carry us productively.

Related event: Anthropic Researcher Argues AI Defies All Existing Analogies(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →