Earlier quoted context omitted.
The problem isn't computers being considered an "other", but rather that other people are using computers to unilaterally scale up the implementation of their own negligence/biases/"policy" while also insulating themselves from corrective feedback or other repercussions. This makes poor results feel willful rather than being considered honest mistakes.
The first part isnt really an assertion of fact but of perspective. Consider if the delivery logistics tracking was executed by men working with paper in an office in the 1920's, and the policy was a managers. That is still "unilateral" as people who are not the manager dont have control over the policy, nor necessarily the capital to create a competitor policy. Additionally, there is corrective feedback. This errant…
Your second part is channeling the efficient market fallacy, and then tautologically writing off "small" problems as not important enough. But once again, the problem is the scale itself. Something that hurts 0.01% of customers/users is never going to move the needle of organizational feedback, but at a scale of ten million customers/users that is still 1000 people that get hurt. Human scale limits fan out and allows direct feedback, surveillance industry scale does neither.