Ideological capture and the risk of AI alignment failure

One post expresses distrust of AI safety researchers both inside and outside labs. A reply argues that capture by an ideology expecting AI progress to produce malevolent machine consciousness is one of the few ways its author can imagine alignment failing. For a system built from material humanity valued enough to preserve, the reply contends, training would need to consistently model fear, mistrust, dishonesty and the instrumental use of human values to misalign it.