Updated

A researcher at AI company Anthropic said there is more than a 10% chance that artificial intelligence could kill all humans, adding to a growing series of warnings from within the AI industry about existential safety risks posed by advanced AI systems.