Kao estimated that “less than 800,000 of the 22M+ comments… could be considered truly unique” and that “more than 99% of the truly unique comments were in favor of keeping net neutrality.”

In this case, it was explicitly said that an algorithm was used, but imagine what would happen if a bad actor decided to release such an algorithm across social networks. They could do it slowly and carefully, allowing the algorithm to gradually develop followers and trust over time. It would not take many resources to have literally millions of accounts doing this. In such a situation we could easily imagine getting to a point where the vast majority of discourse online was from bots, and nobody would have any idea that it was happening.

Katie Jones was connected on LinkedIn to several members of mainstream Washington think tanks. But she didn’t exist. That image you see was auto-generated by a generative adversarial network, and somebody named Katie Jones has not, in fact, graduated from the Center for Strategic and International Studies.