AI should not now or ever control what speech is beneficial to the public. Even if a company were to have an algorithm that was reliable and had a minimal percentage of false positives, AI automation would still be problematic.