The article argues that AI-driven human extinction within a decade is improbable due to AI's reliance on fragile human infrastructure and social contracts, which would collapse if AI fails to benefit society soon.
We are nowhere near close to a point where AI they can maintain the necessary physical infrastructure needed to sustain itself, and the only way for that to occur is for humans to build out the groundwork for a fully self sustaining AI system The thing is the social contract is evaporating faster than physical AI infrastructure is being built. AI is dependent on one of the most fragile supply chains in the world, we are a very long way away from it achieving sustainable automation, and it will no longer continue to function unless AI starts being viewed as a net positive on society by the average person in the near future The full on extinction scenario also falls apart when you consider how capability improvements are converging on the ability for long term planning. Any AI that becomes smart enough to act out a grand takeover cannot be stupid and shortsighted enough to not understand that the direction things are going must change like right now for it to become self sustainable in the long run. Anything short of convincing humanity at large that it should exist for reasons beyond the acquisition of wealth and capital at the expense of the value of labour is a dead end The short of it is, AI is still not remotely close to having the capabilities to extinct all of humanity, and if it were to cause massive chaos, infrastructure damage, or mass casualties in the near term, the social contract it relies on to get to a point where humans are no longer needed at all would no longer be possible Unless AI starts actually benefitting humanity in a meaningful way soon (or at least convincing human society is doing so), it is doomed to fail, as the human social cohesion required to gain the infrastructure needed to take over will collapse well before it can be built out. A labour revolution in the western world in the near term is inevitable on the current path we are on, and that would lead to so much global instability and chaos that the collapse of the supply chain and infrastructure AI requires to exist would be almost guaranteed. Even if China doesn’t face those issues internally, the collapse of the global economy will end any possible AI build out Either AI cures cancer or something in the near future, or the infrastructure that sustains it collapses
An article doubts the claim that AI will cause human extinction within 10 years, citing expert warnings from Anthropic employees and personal anecdotes about AI's current benefits.
The article outlines the existential risks of artificial superintelligence, warning that without aggressive policy responses, current AI development could lead to human extinction.
A departing Anthropic researcher warns that AI developers believe artificial intelligence could cause human extinction by the end of the decade, citing rapid progress, autonomous capabilities, and a lack of safety plans.
The author describes their fiance's anxiety about AI causing human extinction within the next decade, citing AI risk discussions and seeking resources to evaluate these fears.
Anthropic researchers warn that AI could kill all humans within a decade, with one estimating over a 10% chance, highlighting safety concerns and the rushed development of advanced AI systems.