AI Risk Comparisons to Nuclear Weapons Exaggerated; Focus Needed on Current Harms

Comparisons between AI and nuclear risks are often exaggerated. Current AI systems pose real harms that warrant regulation now, not just hypothetical existential threats.

AI labs resist regulation but invoke nuclear-level comparisons, wanting exclusive licensing but calling current systems low risk. This is contradictory.

Nuclear safety principles say failures compound. So we must address current AI harms to prepare for emergent risks, whether AGI or otherwise.

Pushing narratives like AI extinction distracts from regulating current harms. EU AI Act already weakened after lobbying.

No basis exists for if/when AGI emerges. Nuclear risks are scientifically studied. AI extinction is hypothetical, but distracts regulation.