The question is not whether AI can replace us but what role we play in ensuring that AI does not irreversibly harm humanity.