Embed Notice

@ai6yr when people confidently assert that fears of AI causing a mass extinction event are overblown, I think they’re overlooking this sort of thing as an extremely plausible scenario.

The risk of hyper-competent AI systems autonomously wiping us out seems low, but the risk of incompetent human operators making catastrophically bad choices based on the *assumption* that AI systems are hyper-competent seems to be growing day by day.