All of these complaints are about humans making user-hostile decisions, not AI. And software quality is not a reflection of its code quality. Never has been.