AI harbours evil, because unskilled people tend to trust it blindly. People have already been evicted, arrested and harassed by police simply because they choose to trust technology that flags or "recognizes" them, with no proof that they can trust it. This happens automatically. Thus, AI should be treated as potentially malicious, especially when it is sold as neutral.
You pretend that you don't understand the point. Just like Nazi collaborators "pretending they just follow orders" wouldn't see the "bad thing in just making trains go to their destination".
